Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2017-02-08

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

All times shown according to UTC.

Time Nick Message
00:00 nettoweb joined #git
00:01 Dil3mm4 Hello.
00:01 gitinfo Dil3mm4: 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.
00:02 theoceaniscool joined #git
00:02 gopar joined #git
00:02 Dil3mm4 I'm trying to git lfs push some items but I get a msg about the fail of the operation
00:02 minibar hi
00:02 minibar trying to sign commits with git
00:02 Dil3mm4 I've just removed git cache
00:03 Dil3mm4 and added the whole dire to git lfs track
00:03 Dil3mm4 with git lfs track "dir/*"
00:03 Dil3mm4 then git push -u origin master
00:03 Tobbi joined #git
00:03 Dil3mm4 and I get Git LFS: (0 of 0 files, 72 skipped) 0 B / 0 B, 288.88 MB skipped
00:03 Dil3mm4 Even if those 72  files are not online yet
00:03 minibar https://help.github.com/articles/telling-git-about-your-gpg-key/ here it says I only have to tell git about my key ID if my key has a different identity than my git user. both my key and git user have the same email address. shouldn't this be enough for git to identify my key?
00:04 jstimm joined #git
00:04 jstein_ joined #git
00:07 Derperperd joined #git
00:09 thib joined #git
00:11 gopar joined #git
00:11 nowhereman joined #git
00:14 menip joined #git
00:16 howitdo joined #git
00:17 BackEndCoder joined #git
00:22 nevodka joined #git
00:24 dendazen joined #git
00:27 jdelreal joined #git
00:29 jstimm joined #git
00:30 minn joined #git
00:31 jdelreal joined #git
00:32 done joined #git
00:34 dil3mm4 joined #git
00:35 cdown_ joined #git
00:36 shinnya joined #git
00:38 roelmonnens joined #git
00:42 cdown_ joined #git
00:44 pur3eval joined #git
00:47 sarp joined #git
00:51 xissburg joined #git
00:53 MattMake_ joined #git
00:56 pandeiro joined #git
01:01 Gsham joined #git
01:06 matoro joined #git
01:08 Atrumx joined #git
01:08 qpdb joined #git
01:14 pietv joined #git
01:15 pks joined #git
01:20 staafl joined #git
01:22 Guest9 joined #git
01:23 cdg joined #git
01:23 zacts joined #git
01:24 sarp joined #git
01:25 jimi_ joined #git
01:27 Sasazuka joined #git
01:29 xaviergmail joined #git
01:29 zivester joined #git
01:32 jstimm joined #git
01:34 roelmonnens joined #git
01:34 dreiss joined #git
01:34 moei joined #git
01:36 holodoc joined #git
01:36 zzz joined #git
01:40 Sasazuka joined #git
01:42 brent__ joined #git
01:43 cqi joined #git
01:43 pietv joined #git
01:45 DolpheenDream joined #git
01:45 JustJohn joined #git
01:45 pur3eval joined #git
01:46 gkatsev joined #git
01:47 JustJohn left #git
01:49 Derperperd joined #git
01:49 cdg joined #git
01:49 gugah joined #git
01:54 MattMaker joined #git
01:55 plos joined #git
01:59 mischat joined #git
02:01 brian_penguin joined #git
02:02 johnny56 joined #git
02:04 thiago joined #git
02:07 Guest9 joined #git
02:13 gkatsev joined #git
02:16 UniFreak joined #git
02:17 miczac joined #git
02:17 Kaisyu joined #git
02:18 durham joined #git
02:19 finalbeta joined #git
02:23 dsantiago joined #git
02:24 Vampire0 joined #git
02:26 k_sze[work] joined #git
02:27 k_sze[work] joined #git
02:28 Wulf4 joined #git
02:28 jimi_ joined #git
02:30 jameser joined #git
02:35 CyberNetWrk joined #git
02:35 CyberNetWrk left #git
02:37 nd_ joined #git
02:46 durham joined #git
02:46 matoro joined #git
02:52 durham joined #git
02:56 min joined #git
02:56 jj- joined #git
02:57 aavrug joined #git
02:57 ojdo joined #git
02:58 Saturn812 joined #git
02:58 pks joined #git
02:59 Darren__ joined #git
02:59 t-mart joined #git
02:59 d^sh joined #git
03:00 watabou joined #git
03:01 XenophonF joined #git
03:02 hvxgr joined #git
03:06 fstd_ joined #git
03:09 fahadash joined #git
03:13 duderonomy joined #git
03:17 Vortex34 joined #git
03:18 rgrinberg joined #git
03:20 vdv joined #git
03:22 chachasmooth joined #git
03:22 shgysk8zer0_ joined #git
03:28 mminn joined #git
03:28 howitdo joined #git
03:28 howitdo joined #git
03:28 cqi joined #git
03:32 Guest24481 joined #git
03:34 pur3eval joined #git
03:38 kyan Hi... what's this mean, and how can I fix it? It doesn't seem to give much in the way of hints for fixing the problem... https://gist.github.com/ethus3h/87c5fbb36fdd05a2ab6cb9997dd13168
03:38 kyan I assume that it doesn't like something about my repo, but I don't know how to track down the probelm
03:40 streetwitch joined #git
03:40 min joined #git
03:40 kyan Oh! Sorry to bother, solved it.
03:41 sunri5e joined #git
03:41 streetwitch when I do a git pull on my local directory I get an "already up to date" message
03:41 streetwitch But I know it isn't because I just changed it
03:41 streetwitch ideas?
03:41 kyan (To fix it, I created a new branch, checked out the new branch, and then checked out master again and it was good)
03:41 kyan streetwitch: well, git pull only looks at the commits that origin has that your copy doesn't
03:41 kyan git status will show you the state of your working directory
03:42 streetwitch Ah yes
03:42 streetwitch but how do I fix it back to the way it was?
03:43 streetwitch I could git clone again
03:43 kyan Like, to discard your work?
03:43 streetwitch Yeah
03:43 kyan If you don't want to keep your work, I think it's like git reset --hard master or something, but let me look that up first
03:43 streetwitch ty
03:44 kyan streetwitch, have you made any commits that you want to get rid of?
03:44 kyan (or do you just want to get rid of uncommited changes?)
03:45 streetwitch No I haven't
03:45 kyan cool, just do: git reset --hard
03:45 kyan should fix it. (I disclaim any responsibility for things blowing up)
03:45 GodGinrai lol
03:46 streetwitch HEAD is now at 2cd5f68 Desc
03:46 streetwitch ^thats all I get?
03:46 amagawdd joined #git
03:46 kyan Try running git status?
03:46 kyan it should show that working directory is clean
03:46 streetwitch No changes
03:46 streetwitch nope
03:46 kasual joined #git
03:46 streetwitch Do I need to do: git reset --hard master?
03:47 GodGinrai streetwitch: yea, git-reset moves the HEAD pointer to whatever commit you specify
03:47 streetwitch So git pull then?
03:47 streetwitch or what?
03:47 GodGinrai and if you use --hard, it makes the filesystem match the state at that commit
03:47 kyan streetwitch, listen to them, they know more than I do ;)
03:47 GodGinrai streetwitch: "no changes" means that your working directory is clean
03:47 Goplat joined #git
03:47 streetwitch No it doesnt say that tho
03:48 streetwitch modified:   Mage (modified content)
03:48 streetwitch that's what status shows
03:48 kyan Oh! You mean there were no changes from what git status was showing before
03:48 GodGinrai whaand that's after using git-reset?
03:48 GodGinrai *and
03:48 streetwitch Yes
03:49 GodGinrai well
03:49 GodGinrai you could abandon changes on a per file basis with `git checkout HEAD -- filename`
03:49 hexagoxel joined #git
03:49 streetwitch Is there a disadvantage to just cloning the whole thing again?
03:50 streetwitch because this is taking way to long to figure out
03:50 kyan if it's a 2gb repo and you have dial up internet, then yes there's a disadvantage ;3
03:50 GodGinrai streetwitch: the disadvantage is that you won't learn how to fix this problem when recloning is not an option
03:51 kyan I still don't get why reset --hard didn't work
03:51 minn joined #git
03:52 GodGinrai kyan: no idea.  Either he's not telling us something, or git-reset doesn't default to HEAD if not specified
03:53 kyan Oh, hmm. I guess I assumed it would
03:53 * kyan should not make assumptions -_-;
03:53 GodGinrai streetwitch: roundabout way to do it, but you could also do this: `git stash; git stash drop`
03:53 GodGinrai that should clean your working directory
03:54 GodGinrai of changes
03:54 streetwitch I just cloned
03:54 streetwitch sorry
03:58 kyan I'm still getting that assertion failure :(
04:00 mischat joined #git
04:00 k_sze[work] joined #git
04:02 min joined #git
04:02 scarabx joined #git
04:06 afuentes joined #git
04:10 dweelie joined #git
04:10 kyan Ah! think I may have found it now, I'd acidentally deleted the .git file from one of my submodules
04:11 a_thakur joined #git
04:11 minn joined #git
04:12 a_thakur joined #git
04:15 kyan Yup, that fixed it ????
04:16 mischat joined #git
04:24 vacho joined #git
04:28 k_sze[work] joined #git
04:30 askb joined #git
04:31 LiftLeft joined #git
04:31 navidr joined #git
04:31 robotroll joined #git
04:32 mehola joined #git
04:33 LiftLeft what would be the easiest way to move files in a local git directory to another local git directory, in bash and I want to respect .gitignore?
04:33 chachasmooth joined #git
04:34 sargentmki joined #git
04:35 LiftLeft I want to create a complete copy of the latest commit on directory "A" and clone it over to directory "B"
04:37 amagawdd joined #git
04:38 lucasem joined #git
04:38 sbulage joined #git
04:39 Cabanoss- joined #git
04:39 govg joined #git
04:42 min joined #git
04:51 MACscr joined #git
04:53 Emperor_Earth joined #git
04:54 Zarthus I can think of one hacky way, bar re-cloning.
04:54 xall joined #git
04:56 Zarthus I think just a local fs clone will suffice, though.
04:56 Zarthus I don't have a good copy currently accessable to me to verify.
04:57 LiftLeft I need to do it with bash
04:59 Zarthus LiftLeft: `git clone a b` is bash, surely? :P
04:59 Zarthus unless you wish to imply git is not installed on your system.
05:00 LiftLeft it's installed
05:00 Zarthus Then I'd give `git clone a b` a shot. I'm not entirely confident if it strips out gitignored files.
05:03 ayogi joined #git
05:04 solenodic joined #git
05:04 LiftLeft Zarthus: I need to copy the commit into already existing repo
05:04 Zarthus you're just talking about one commit?
05:05 pur3eval joined #git
05:05 danimal2 joined #git
05:05 LiftLeft yes, the latest
05:07 Zarthus you could git push to the remote on your local filesystem, then pull. Or use git-format-path
05:07 Zarthus patch
05:08 LiftLeft so I could just pull from repo "A" to repo "B"?
05:13 LiftLeft I think I had a bad idea
05:13 Zarthus you have two repositories in your filesystem, I doubt anything stops you from pushing from one to the other
05:18 m4sk1n joined #git
05:20 xall joined #git
05:24 justan0theruser joined #git
05:27 mischat joined #git
05:28 dtype Zarthus: assuming file permissions allow it, works great
05:29 aidalgol joined #git
05:29 ac3takwas joined #git
05:32 phroa LiftLeft: you could potentially pass --work-tree=/path/to/thing to 'git checkout thecommit', but that's "thought experiment" territory for me and deserves a !backup
05:32 gitinfo LiftLeft: Worried about your data while trying stuff out in your repo? The repository in its entirety lives inside the .git directory in the root of your work tree so to backup everything `cp -a path/to/workdir path/to/backup` or equivalent will suffice as long as the repo is not modified during backup. See also http://sethrobertson.github.com/GitBestPractices/#backups
05:34 LiftLeft I'm changing I'm going to do stuff
05:34 LiftLeft I don't need that problem to be solved anymore
05:39 xall joined #git
05:40 howitdo joined #git
05:40 howitdo joined #git
05:41 an0ther joined #git
05:41 an0ther So I have some merge bubbles in my commit history, is there a way to get rid of them?
05:42 an0ther The problem is that they are several commits back.
05:42 Vampire0_ joined #git
05:42 _ikke_ an0ther: Have these merges been pushed yet?
05:42 an0ther Nope :)
05:43 _ikke_ rebase can flatten history
05:44 hahuang62 joined #git
05:45 an0ther _ikke_: I tried that, but how do I connect it to the several commits after?
05:52 brent__ joined #git
05:53 FabTG joined #git
05:58 scarabx left #git
06:00 redhedded1 joined #git
06:01 freimatz joined #git
06:03 MineCoins joined #git
06:03 an0ther _ikke_: The problem is I have commits after the merge of two branches. So how do I rebase? The history of the two branches is already intertwined
06:04 rscata joined #git
06:05 pur3eval joined #git
06:06 _ikke_ an0ther: You have to do multiple rebases then
06:07 _ikke_ !situation
06:07 gitinfo Please pastebin the relevant output of `git log -n 20 --all --graph --format="%h %p %d"` to give us an idea about what your situation is (increase the -n argument if your problem is more complex than 20 commits can explain)
06:09 Gsham joined #git
06:11 an0ther https://gist.github.com/anonymous/c920989353bcadf473c332c854822e30
06:14 jameser joined #git
06:15 _ikke_ an0ther: You can try git rebase -i 18eeca3 experiment
06:15 _ikke_ if that turns out okay, then you can reset master to experiment
06:15 Topic for #git is now Welcome to #git, the place for git help and a bag of wheeee | 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 | Are you even reading this? Prove it by submitting to https://madeitwor.se/git-jokes
06:16 _ikke_ Eugene: You keep trying :P
06:16 dec0n joined #git
06:16 Eugene Also, the *.madeitwor.se service is no longer being maintained because StartSSL got revoked as a CA and I'm not paying $200 for a wildcard certificate
06:16 Eugene Custom subdomains available by request(I need to set up SSL)
06:16 netj joined #git
06:17 k_sze[work] joined #git
06:17 Vampire0 joined #git
06:17 bocaneri joined #git
06:18 an0ther Eugene: Not that I know much about CA's, but can't you use Let's Encrypt?
06:18 k_sze[work] joined #git
06:18 _ikke_ There are extensions to create LE subdomain certs on the fly
06:20 _ikke_ Eugene: https://github.com/EugeneKay/agile-php lol
06:20 _ikke_ very bare-bones
06:20 Eugene I am using LE, but I dont' want to recompile nginx for that module
06:21 Eugene And yes, thatsthejoke.jpg
06:22 minn joined #git
06:29 waveclaw joined #git
06:30 Vampire0 joined #git
06:30 an0ther So any ideas on how to flatten those merge bubbles? Or is it not possible now that I've added commits after the merge?
06:31 PaulCapestany joined #git
06:32 rominronin joined #git
06:32 Guest82 joined #git
06:34 Vampire0_ joined #git
06:35 GodGinrai joined #git
06:37 sbulage joined #git
06:37 qt-x joined #git
06:39 Nokurn joined #git
06:39 an0ther anyone? anyone? Bueller?
06:40 bocaneri joined #git
06:43 ojacobson an0ther: LE's policy on wildcards is "no", too
06:44 svm_invictvs joined #git
06:44 ojacobson What's https://gist.github.com/anonymous/c920989353bcadf473c332c854822e30 the output of?
06:44 rominron_ joined #git
06:45 iamtakingiteasy joined #git
06:45 abra0 joined #git
06:46 bocaneri joined #git
06:47 an0ther ojacobson: git log --all --graph --format="%h %p %d"
06:48 ojacobson Thanks
06:49 Eugene I thought LE said "yes, but we don't know how yet"
06:49 Vampire0 joined #git
06:49 Eugene Innany c
06:49 jfr joined #git
06:49 Eugene In any case, I don't think anybody ever used it.
06:50 Eugene But you should use https://vomitbin
06:50 iamtakingiteasy joined #git
06:50 Eugene Https://vomitb.in thanks autocorrect
06:50 ojacobson You could flatten them by rebasing on top of 18eeca3
06:50 ojacobson but don't
06:50 ojacobson they look fine
06:51 ojacobson flattening the twigs will have the effect of rewriting most or all of the commits since 18eeca3
06:52 an0ther ojacobson: Yeah, that'd be okay, I can confirm the files are the same by diffing against this master, I was curious how it could be done
06:52 dersand joined #git
06:52 ojacobson Eugene: "can't" and "won't" have the same semantics to a project outsider ;)
06:52 bket joined #git
06:53 hahuang62 joined #git
06:53 Eugene People are dumb.
06:55 osse so am i
06:55 Eugene LE doesn't understand RFC2544 addresses, so I had to disable IPv4 on my home fileservet.
06:55 Eugene Its a really dumb problem and I hate that I even have it.
06:55 ThomasLocke joined #git
06:55 ThomasLocke joined #git
06:55 ojacobson osse: people
06:55 ojacobson ?
06:57 Grum joined #git
07:00 _ng joined #git
07:00 dreiss joined #git
07:02 Macaveli joined #git
07:02 ihradek joined #git
07:02 ThomasLocke_ joined #git
07:02 ThomasLocke_ joined #git
07:04 Raging_Hog joined #git
07:05 nidr0x joined #git
07:07 algun joined #git
07:09 zeroed joined #git
07:09 zeroed joined #git
07:09 zacts joined #git
07:14 roelmonnens joined #git
07:14 chele joined #git
07:15 roelmonnens joined #git
07:16 JeroenT joined #git
07:16 spacelord_ joined #git
07:17 MrMojit0 joined #git
07:17 minn If I wanted to see every commit affecting a file up to a particular release, the command to build off would be 'git log --follow <release> <file>', correct?
07:18 Vampire0_ joined #git
07:19 curVV joined #git
07:19 ozmage joined #git
07:21 freimatz joined #git
07:24 curVV joined #git
07:24 curVV hi, when branching from develop to release, doing some release specific tweaking and then merging (with --no-ff) this release branch back into both main and develop branches, creates these two different merge commits. what would happen with the next release? would it be a problem even though there are no difference in files? would the two (identical) commits then exist in both master and develop?
07:25 curVV looking at this approach here: http://nvie.com/posts/a-successful-git-branching-model/
07:27 elect_ joined #git
07:28 ToxicFrog joined #git
07:29 osse minn: yes
07:29 minn osse: Cool, thanks
07:30 osse minn: the --follow is only needed if the files has been renamed.
07:31 Skitzo_ joined #git
07:31 zefferno joined #git
07:32 spacelord_ joined #git
07:34 Mikerhinos joined #git
07:36 jameser joined #git
07:36 minn osse: It's working as intended, then, since I need to follow renames :) Thanks
07:38 spacelord_ joined #git
07:39 keepguessing joined #git
07:44 roelmonn_ joined #git
07:46 keepguessing I have moved to a new team and here their project has 3 branches, master zone1 zone2. Each branch has a directory calledgenerated where the changes are different.
07:47 keepguessing now for changes where the generated is not involved the team does a git merge between the two branches and for the changes which involves generated they use cherrypick
07:47 mda1 joined #git
07:48 keepguessing I have not used this pattern of branching before. I believe that in case 1 changes should be cherry-picked and in case two they should be committed.
07:48 keepguessing Any suggestions?
07:48 jast testing, ignore me... git:github:git/git:8a8f1210
07:48 gitinfo Git web link: https://github.com/git/git/commits/8a8f1210
07:50 jast keepguessing: I'm not really understanding the point of what they're doing
07:51 keepguessing jast: they use http://stackoverflow.com/questions/1838873/visualizing-branch-topology-in-git/34467298#34467298 to visualize the branches
07:52 keepguessing jast: they say that this way they know the convergence of the code. I seriously do not understand the reason for doing this.
07:53 jast "convergence of the code"? that could mean pretty much anything...
07:54 keepguessing jast: the aliases describe in the stack overflow above shows "arrow marks" of the various branches merging.
07:54 pur3eval joined #git
07:54 keepguessing jast: and that somehow tells them something which I cannot understand.
07:54 netj joined #git
07:55 jast yeah, that's what merges look like
07:56 jast basically the point of merges is to leave an explicit record of "person X combined these two pieces of diverged history at this point in time", which can be useful to know later on. you don't get the same information if you cherry-pick. but that only explains why they merge, not why they don't. :)
07:56 keepguessing jast: can you understand what they are trying to achieve. I asked and they look at me strangely.
07:57 keepguessing jast I see.
07:57 Skitzo_ joined #git
07:57 jast plus, for example, you can have a more high-level look at the history, leaving out the commits that got merged in: git log --first-parent
07:57 keepguessing jast: but wont you see all the refs of one branch in the other ?
07:58 jast what do you mean by 'refs'?
07:58 keepguessing hash
07:58 jast okay. ref, in Git, means "reference to a commit", i.e. a hash, branch name, etc.
07:58 jast (refs can point to tag objects, too, but that's a minor detail)
07:59 jast and yes, you will, but the structure of the history is such that you can still tell which side of the merge each commit is in
07:59 jast the default view just sorts everything by date, but there are many other ways to format the history
07:59 keepguessing ok wont merging also merge all the commits? and show those commits in the merge branches history?
07:59 jast yes, it will
08:00 keepguessing jast ok so if I do a git log then I would see the other branches commit that I merged.
08:00 jast yes
08:00 keepguessing jast: what I backout that ref assuming it was in this branch?
08:00 keepguessing s/ref/commit/
08:01 jast "back out"?
08:01 keepguessing s/what /what if/
08:01 grawity jast: hg &c. term for `git revert`
08:01 BSAlb joined #git
08:01 Skitzo_ Hey guys
08:01 Skitzo_ I have a problem with deferred deep link with facebook at iOS, The problem is very very very ...... very weird.
08:01 Skitzo_ My step is below,
08:01 Skitzo_ 1, I create a AD and preview at the iPhone's facebook app.
08:01 Skitzo_ 2, I use facebook native app view this ad, and click the install now button, there will open the App Store show my app detail.
08:01 Skitzo_ 3, I use Xcode install my app, and app launch. Attention please,  at this
08:01 Skitzo_ step, if I use difference facebook accounts,  there will have difference results:
08:01 gitinfo Skitzo_: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.
08:01 Skitzo_ result √:  if I use my account , the new install app will received the right deffered deep link url and no error. It's very nice!
08:01 Skitzo_ result ×: but if I use other's accounts, the new install app wouldn't received anything, no url and no error too.
08:01 Skitzo_ I test many times, always like this. why why why?
08:01 cbreak no spamming please
08:02 jast keepguessing: so you want to know what happens if you revert a merge commit?
08:02 osse someone pressed Ctrl-V in the wrong place
08:02 Skitzo_ Sry, ..
08:02 osse or rather Cmd-V
08:02 grawity this isn't facebook tech support you know
08:02 jast Skitzo_: hi, I can tell you things about Git ;)
08:02 cbreak Skitzo_: I recommend against using facebook
08:02 keepguessing grawity: actually no. i have not used hg ... jast yes
08:02 cbreak it's garbage
08:03 cbreak last I read, they publish lies as news, without even bothering to validate them.
08:03 jast okay, well, generally speaking, there are multiple things you could do that might qualify as "backing out"
08:03 keepguessing grawity: my company used to use perfore. may be comes from there. but yeah right intution
08:03 keepguessing jast: yeah revert is what I mean here.
08:03 jast "git revert" specifically creates a commit out of the reverse of the diff some other commit introduces
08:03 jast with merge commits that's a bit special
08:03 MattMaker joined #git
08:04 jast revert will ask you specifically which branch you want to use as the base for generating the reverse diff (most of the time you'll want the first parent, i.e. -m 1)
08:04 jameser joined #git
08:04 jast but the commits from the merge will stay in the history, because revert does not actually undo the merge, just the diff it results in
08:04 jast so you'll still see them in your history, and repeating the same merge later will have no effect because, as far as Git is concerned, the commits are already merged
08:05 keepguessing jast I am not attempting to revert the merge commit. since the commits from branch1 get shown in branch2 after the merge
08:05 curVV_ joined #git
08:05 Skitzo_ cbreak:  all media is lie... got soe news? I need help with this crap..
08:05 jast this is why it's a very good idea to be very sure about your merge before you push/publish it :)
08:05 cbreak revert does not change existing history, it merely adds new history that undoes previous changes
08:05 cbreak reset / rebase / ... allows changing existing history, replacing it with new one.
08:05 Skitzo_ http://pastebin.com/SxMt9vKB if someone can help me with this, should be cool
08:06 cbreak Skitzo_: use C++
08:06 keepguessing ah ok got it. so there is no side effect or bad practice here by the team.
08:06 jast urgh, c++
08:06 keepguessing jast: they might be doing it for the genuine reason maybe.
08:06 Timvde joined #git
08:06 jast keepguessing: well, I don't know where exactly their distinction between using merge and cherry-pick is coming from, but then again I don't know what's so special about that one subdir in the repo :)
08:08 keepguessing jast: for example its like "certs" in that directory are different in each branch.
08:08 Alienpruts joined #git
08:08 cbreak that sounds bad
08:08 cbreak keepguessing: repositories should not contain such things
08:08 cbreak it'll make merging a pain in the ass
08:08 keepguessing cbreak: yes it does not I just gave an example. Wanted it to be obvious
08:09 bill99 joined #git
08:11 Vampire0_ joined #git
08:12 Tobbi joined #git
08:13 matoro joined #git
08:14 a_thakur joined #git
08:16 ssardana08 joined #git
08:18 jnavila joined #git
08:18 theoceaniscool joined #git
08:20 andream joined #git
08:24 martinklepsch joined #git
08:25 MeltedLux joined #git
08:27 schneider joined #git
08:27 djb-irc joined #git
08:27 andream joined #git
08:28 JeroenT joined #git
08:28 MapMan joined #git
08:28 CtrlC joined #git
08:28 CtrlC joined #git
08:29 tabakhase joined #git
08:31 jdiez joined #git
08:32 cdown joined #git
08:32 thierryp joined #git
08:33 a_thakur joined #git
08:34 Dumblez0r joined #git
08:35 Sigma joined #git
08:35 j7k6 joined #git
08:35 Sc0rpius joined #git
08:36 GodGinrai joined #git
08:36 Someguy123 joined #git
08:36 King_Hual joined #git
08:36 King_Hual joined #git
08:37 DaveTaboola joined #git
08:38 andream joined #git
08:39 JeroenT joined #git
08:39 minn Is there any way to specify that newlines be escaped in message bodies when using %B or %b as a format specifier? Roughly, I either want to capture the message body in a single line to ensure data integrity while parsing commit logs.
08:40 nanga joined #git
08:40 nanga joined #git
08:41 minn The 'either' is a typo, whoops.
08:41 grumble joined #git
08:41 jast minn: if you're parsing, you may want to use tformat instead of format, to have the commits separated by NUL bytes, so then it'd be easy to detect the boundaries without escaping anything else. I don't think there's anything built in to escape newlines, anyway...
08:42 jast actually I didn't mean tformat, misremembered that
08:43 jast -z
08:43 Tobbi joined #git
08:44 Balliad joined #git
08:44 ThomasLocke_ joined #git
08:44 ThomasLocke_ joined #git
08:45 minn jast: That's slightly unfortunate, but I suppose it's just an annoyance rather than an actual problem. Thanks for the recommendation about -z :)
08:45 PCatinean joined #git
08:45 Nicmavr joined #git
08:45 jast it's actually easier dealing with that than doing unescaping, if you ask me
08:46 kurkale6ka joined #git
08:46 djb-irc joined #git
08:47 zack_s_ joined #git
08:50 ruskie joined #git
08:50 zack_s_ somehow I accidentally deleted a commit from my repo remotely, it was not the last commit
08:50 _ikke_ What did you do?
08:50 zack_s_ do I have any problems now when I sync with the repo locally?
08:51 zack_s_ I wanted to delete a tag
08:51 ihradek joined #git
08:51 zack_s_ I have also an svn sync plugin which keeps an svn version of the repo in sync with the git repo
08:51 mikecmpbll joined #git
08:51 minn jast: True. Since messages can't contain the null character, that actually makes parsing messages pretty easy.
08:51 navidr joined #git
08:55 andream joined #git
08:56 MatthewAllan93 joined #git
08:56 MatthewAllan93 joined #git
08:58 Skitzo_ !needsupport  Require: skillz with ios SDK facebook...
08:59 jast Skitzo_: try somewhere that isn't a Git support channel?
08:59 andream joined #git
08:59 Skitzo_ Tryed 101 chans.. error 404 not found
09:00 jast like, you know, stackoverflow...
09:00 a_thakur joined #git
09:01 _ikke_ zack_s_: can you post a !transcript somewhere?
09:01 gitinfo zack_s_: 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
09:02 mikecmpbll joined #git
09:02 bongjovi joined #git
09:04 _ikke_ Skitzo_: direct messaging is not going to help you
09:04 Skitzo_ _ikke_:  i c you
09:06 borkr joined #git
09:06 j7k6 joined #git
09:06 ExeciN joined #git
09:06 iamthad joined #git
09:07 freezer joined #git
09:08 t-mart_ joined #git
09:12 garethdaine joined #git
09:13 _ADN_ joined #git
09:13 nettoweb joined #git
09:16 ThomasLocke_ joined #git
09:16 ThomasLocke_ joined #git
09:17 marcogmonteiro joined #git
09:17 Shashikant86 joined #git
09:17 t-mart joined #git
09:19 JeroenT joined #git
09:24 andream left #git
09:24 manuelschneid3r joined #git
09:25 rorro joined #git
09:26 DieguezZ joined #git
09:26 andream joined #git
09:27 zeroed joined #git
09:28 andream left #git
09:28 solenodic joined #git
09:32 plos joined #git
09:35 sven^ joined #git
09:35 sven^ joined #git
09:35 rewt`` joined #git
09:36 tang^_ joined #git
09:36 rchavik joined #git
09:36 mehola_ joined #git
09:37 a_thakur_ joined #git
09:37 unbalanced joined #git
09:37 dfas` joined #git
09:37 qt-x1 joined #git
09:38 Serus_ joined #git
09:38 ChristopheT joined #git
09:38 denisMone joined #git
09:38 danmackay_ joined #git
09:38 PtxDK joined #git
09:39 lurke joined #git
09:39 Serpent7776 joined #git
09:39 snowinferno_ joined #git
09:39 Danzeel joined #git
09:40 ekneuss_ joined #git
09:40 Pieter_ joined #git
09:40 MarkusDB1 joined #git
09:40 runlevl4_ joined #git
09:40 curVV joined #git
09:40 Macaveli_ joined #git
09:40 localh0s1 joined #git
09:40 dwts_ joined #git
09:40 hvxgr_ joined #git
09:40 orzo_ joined #git
09:40 marcadde1 joined #git
09:40 ikonia_ joined #git
09:40 jbondhus_ joined #git
09:40 reto__ joined #git
09:41 msm joined #git
09:41 iamthad_ joined #git
09:41 dzho_ joined #git
09:41 imMute joined #git
09:41 imMute joined #git
09:41 Mathiasd1 joined #git
09:41 rumble joined #git
09:41 tokkee joined #git
09:41 cjohnson_ joined #git
09:41 jrm2 joined #git
09:41 Peng_ joined #git
09:41 p4trix joined #git
09:42 Ariadeno_ joined #git
09:43 pur3eval joined #git
09:43 reactormonk How do I squash all changes between branch A and branch B into a single commit?
09:44 Pursuit_ joined #git
09:44 JoshL- joined #git
09:44 dan2k3k5 joined #git
09:44 KnightsOfNi_ joined #git
09:44 javaJake_ joined #git
09:44 _Patrick_ joined #git
09:44 AndrewAlexMac_ joined #git
09:44 vishal_ joined #git
09:44 PioneerAxon_ joined #git
09:44 kkirsche joined #git
09:44 marvi_ joined #git
09:44 p53ud0ny- joined #git
09:44 florian^ joined #git
09:44 exounx_ joined #git
09:44 MaekSo_ joined #git
09:44 tterrag|alt joined #git
09:44 jast reactormonk: "between"? do you essentially want to "merge" changes from branch1 into branch2 but squashed?
09:44 preaction_ joined #git
09:44 reactormonk jast, yeah.
09:44 ThiefMasterrr joined #git
09:44 Sc0rpius_ joined #git
09:44 Korvin[BNC] joined #git
09:44 Korvin[BNC] joined #git
09:44 h12o_ joined #git
09:44 caffeinatedcode` joined #git
09:44 bakingbread joined #git
09:44 jast git checkout branch2; git merge --squash branch1
09:44 bouncy- joined #git
09:44 netwoodle joined #git
09:45 Boniche- joined #git
09:45 nellicus joined #git
09:45 TSS3000RO joined #git
09:47 Levex joined #git
09:48 heinrich5991 joined #git
09:49 qguv joined #git
09:50 Shashikant86 joined #git
09:51 jwr joined #git
09:51 Aaeriele joined #git
09:52 yan joined #git
09:52 coasterz joined #git
09:52 askb joined #git
09:52 Jonuz joined #git
09:52 solenodic joined #git
09:52 m1dnight1 joined #git
09:54 monokrome joined #git
09:54 nikow joined #git
09:54 SebastianFlyte joined #git
09:55 ferr1 joined #git
09:56 Shashikant86 joined #git
09:56 Vampire0 joined #git
09:56 ThomasLocke_ joined #git
09:56 nowhereman joined #git
09:57 JeroenT joined #git
09:59 t-mart_ joined #git
10:03 _ikke_ Weird, getting merge conflicts where LOCAL is empty. I would expect git would just be able to merge that...
10:04 _ikke_ hmm
10:04 _ikke_ perhaps because it was explitly remvoed
10:04 _ikke_ removed
10:04 osse _ikke_: i've had that multiple times. I think it's when the context is different
10:04 a3Dman_ joined #git
10:06 achadwick joined #git
10:07 _ikke_ In this case, it's because it was removed in LOCAL and updated in REMOTE
10:08 osse makes sense
10:08 osse like a delete/modify conflict but for a hunk instead of a file :p
10:08 _ikke_ correct
10:10 chll_ joined #git
10:13 troulouliou_div2 joined #git
10:13 diogenese joined #git
10:13 mischat joined #git
10:13 redeemed joined #git
10:18 Shashikant86 joined #git
10:18 ac3takwas joined #git
10:18 mischat_ joined #git
10:20 UniFreak joined #git
10:22 jimsio joined #git
10:23 tvw joined #git
10:27 overlord_tm joined #git
10:28 _ADN_ joined #git
10:31 tvw joined #git
10:35 flaviodesousa joined #git
10:35 ruxu joined #git
10:36 localh0s1 left #git
10:36 GodGinrai joined #git
10:37 localh0st joined #git
10:37 mischat joined #git
10:39 csdcs joined #git
10:39 DieguezZ joined #git
10:43 fk joined #git
10:43 Univ joined #git
10:45 mischat_ joined #git
10:46 irqq joined #git
10:47 Univ JOIN #django
10:47 ThomasLocke_ joined #git
10:47 ThomasLocke_ joined #git
10:50 t-mart joined #git
10:54 geoid_ joined #git
10:57 mehola joined #git
10:58 Darcidride joined #git
10:59 pbrewczynski joined #git
10:59 nowhereman joined #git
11:01 solenodic joined #git
11:08 mckendr joined #git
11:08 tvw joined #git
11:09 ac3takwas joined #git
11:09 Ramen joined #git
11:10 mastro joined #git
11:10 ajd joined #git
11:11 nowhereman joined #git
11:12 dviola joined #git
11:12 khs joined #git
11:12 mmlb joined #git
11:12 ajd if I have a bare/mirror repo which I use as a reference repo for an automated job that runs git clone every now and then, and I happen to have a cronjob to run git fetch on the reference repo, is there any chance of encountering half-written objects or whatever that will cause the clone to fail?
11:14 Vampire0_ joined #git
11:14 nowhereman joined #git
11:17 roelmonnens joined #git
11:21 Snugglebash joined #git
11:22 Anticom joined #git
11:22 khs left #git
11:24 moritz ajd: I think git uses file locking to avoid that
11:26 leeN joined #git
11:27 _ikke_ newer git version have atomic ref updates
11:28 _ikke_ atomic across all refs, ref updates themselves already were atomic
11:32 pur3eval joined #git
11:32 jost__ joined #git
11:32 freimatz joined #git
11:33 OverCoder joined #git
11:33 JeroenT joined #git
11:39 rnsanchez joined #git
11:39 mmlb joined #git
11:39 basiclaser joined #git
11:39 solenodic joined #git
11:40 guest1234 joined #git
11:41 UniFreak joined #git
11:42 Balliad joined #git
11:43 jameser joined #git
11:45 basiclaser joined #git
11:45 marcogmonteiro joined #git
11:46 dviola joined #git
11:48 dendazen joined #git
11:49 JeroenT joined #git
11:49 tlaxkit joined #git
11:52 jameser joined #git
11:52 mischat joined #git
11:53 afuentes joined #git
11:53 ertesx joined #git
11:55 ThomasLocke joined #git
11:55 ac3takwas Hello
11:56 ac3takwas Is this command valid?
11:56 ac3takwas `git add -au`
11:56 ac3takwas I tried to google it,  but there doesn't seem to be any usage for it.
11:56 ac3takwas s/for/it
11:56 ac3takwas s/for/of
11:56 _ikke_ -a and -u seem a bit redundant / mutual exclusive
11:56 _ikke_ sorry
11:57 _ikke_ -a does not exist for add even
12:00 jagob joined #git
12:00 geoid_ joined #git
12:01 muldover joined #git
12:01 Shashikant86 joined #git
12:02 Alienpruts joined #git
12:05 netj joined #git
12:06 finalbeta joined #git
12:07 Seveas joined #git
12:12 quite joined #git
12:12 quite joined #git
12:17 jameser joined #git
12:18 stamina joined #git
12:19 sbulage joined #git
12:20 tcorneli joined #git
12:21 Shashikant86 joined #git
12:22 tcorneli joined #git
12:22 PCatinean_ joined #git
12:22 minn If I'm reading the documentation correctly, 'git ls-tree --full-tree -r <release>' should list every blob and tree object for the given release, but nothing added after that release, correct?
12:24 JeroenT joined #git
12:25 daivyk joined #git
12:27 DaveTaboola joined #git
12:30 theoceaniscool joined #git
12:30 JeroenT joined #git
12:31 nowhereman joined #git
12:32 mingrammer joined #git
12:33 albel727 joined #git
12:33 osse yes
12:35 nedbat minn: i've never had to use ls-tree, what will you do with that?
12:36 minn osse: Cool, thanks.
12:36 osse minn: you keep writing <release>... you mean a specific tag right?
12:37 minn osse: Right, I mean a tag.
12:37 GodGinrai joined #git
12:38 osse minn: good. But btw, the log and ls-tree also work for any branch name or specific commit
12:38 sbulage joined #git
12:38 minn osse: Right, I think I meant reference, but I'm only using it to examine particular tags correspondign to 'releases' (not a git concept).
12:39 YuGiOhJCJ joined #git
12:40 minn nedbat: Attempting to simultaneously compile a dataset and learn a new version control system :)
12:40 nedbat minn: ls-tree is super-advanced, and not needed for day-to-day operations, unless I'm missing something.
12:42 osse ls-tree is a nice to way to produce a list of all files in a commit
12:42 osse "super-advanced" sounds super-advanced :p
12:44 jast it's slightly low-level, but pretty straightforward still
12:45 masuberu joined #git
12:48 Levex joined #git
12:49 tvw joined #git
12:49 ChiLL_SurF joined #git
12:52 cdown joined #git
12:53 UrsoBranco joined #git
12:53 robattila256 joined #git
12:57 overlord_tm joined #git
12:57 styler2go joined #git
12:58 mehola joined #git
12:58 Gloomy joined #git
13:01 Sound joined #git
13:01 boombatower joined #git
13:01 govg joined #git
13:01 Raging_Hog joined #git
13:02 chrysa joined #git
13:03 DolpheenDream joined #git
13:06 JeroenT joined #git
13:06 synthroid joined #git
13:06 nettoweb1 joined #git
13:14 ash_workz joined #git
13:15 Snugglebash joined #git
13:16 miczac joined #git
13:17 jimi_ joined #git
13:17 DieguezZ joined #git
13:20 Snugglebash joined #git
13:20 nowhereman joined #git
13:20 pur3eval joined #git
13:24 roelmonnens joined #git
13:27 jeffreylevesque joined #git
13:27 serialoverflow joined #git
13:27 Emperor_Earth joined #git
13:29 Macaveli joined #git
13:33 zivester joined #git
13:33 buzzLightBeer joined #git
13:34 eroussel joined #git
13:35 manuelschneid3r joined #git
13:35 jophish joined #git
13:36 jophish hi all
13:36 sbulage joined #git
13:36 jophish I'm in my mergetool and I'd like to see the commit messages of the commits in this merge
13:36 jophish what can I do?
13:39 mgedmin open a different shell and use git show (or tig, or git gui, or ...)
13:39 geoid_ joined #git
13:41 Tobbi joined #git
13:42 jophish hmm, git show was what I tried, perhaps it didn't do what I wanted becuase of git-imerge
13:42 jophish thanks mgedmin
13:43 ihradek joined #git
13:43 mgedmin I was thinking you'd provide the commits you wanted to see to git show
13:43 Celelibi Is there a way to programatically tell if the working directory is dirty?
13:44 Celelibi I would like to make a git command for things I usually do.
13:44 Celelibi But the "git stash pop" in the end fail if there wasn't a stash saved because the working directory was clean.
13:45 osse Celelibi: are you writing a script? Consider sourcing git-sh-saetup
13:45 osse setup
13:45 Gsham joined #git
13:45 jophish mgedmin: well, I did --parents
13:46 osse git log HEAD..whatyoumerge
13:46 osse ez
13:46 Gamecubic joined #git
13:47 dendazen joined #git
13:47 ShalokShalom joined #git
13:50 Celelibi osse: that's interesting, but I'm not sure how this would help.
13:50 masuberu joined #git
13:50 osse Celelibi: there are helper functions to make sure the directory is clean, and abort if not
13:50 osse Celelibi: at least you can tell how they've determined it
13:51 marcogmonteiro joined #git
13:51 HardlySeen joined #git
13:52 sharpy joined #git
13:52 sbulage joined #git
13:57 Hudu joined #git
13:59 mattcen joined #git
14:00 mischat_ joined #git
14:00 JeroenT joined #git
14:02 roidal joined #git
14:02 roidal hey
14:04 roidal does the fetch/pull protocol allow to transmit more than one pack-file per request?
14:04 theoceaniscool joined #git
14:06 ToBeCloud joined #git
14:07 Macaveli joined #git
14:07 xdv joined #git
14:10 Raging_Hog joined #git
14:11 Shashikant86 joined #git
14:12 jimi_ joined #git
14:12 ronskisika joined #git
14:14 Vampire0 joined #git
14:14 manuelschneid3r joined #git
14:15 Groovin joined #git
14:17 dhollinger joined #git
14:19 Celelibi osse: I don't understand everything, but it looks like it basically does a git diff --quiet to check if the working directory is clean.
14:20 chipotle joined #git
14:22 tlaxkit joined #git
14:25 nettoweb joined #git
14:26 mischat joined #git
14:27 rivarun joined #git
14:28 ShekharReddy joined #git
14:31 Tobbi joined #git
14:31 Celelibi Any better idea for my command than "rebaseconfigonmodifs"? http://paste.debian.net/913404/
14:34 Shashikant86 joined #git
14:36 GodGinrai joined #git
14:40 wiesel joined #git
14:40 geoid_ joined #git
14:43 Shashikant86 joined #git
14:43 nckpz joined #git
14:44 shinnya joined #git
14:47 re1 joined #git
14:49 rorro joined #git
14:49 al-damiri joined #git
14:50 Scourje joined #git
14:50 Scourje left #git
14:51 mknooihuisen joined #git
14:51 solenodic joined #git
14:53 Levex joined #git
14:53 aLeSD joined #git
14:53 aLeSD hi all
14:53 aLeSD I have a (force updated) running git pull
14:54 aLeSD my git seems update with the remote but other clients don't see my commits
14:59 Vampire0_ joined #git
14:59 mknooihuisen git clone hangs endlessly on “cloning into…” but only on my build server.  Any ideas?
15:00 Dougie187 joined #git
15:00 netj joined #git
15:01 daivyk joined #git
15:01 texinwien_ joined #git
15:01 AaronMT joined #git
15:01 cdown joined #git
15:03 chipotle joined #git
15:04 sunri5e joined #git
15:04 Cassiopaya joined #git
15:04 tobiasvl aLeSD: what do you mean? if you git pull and there's a "(force updated)" message then that usually means you can't merge, but that the remote-tracking branch force updated
15:05 tobiasvl aLeSD: also what does pulling have to do with making other clients see your commits
15:06 babyswizz joined #git
15:08 Gloomy joined #git
15:08 nickabbey joined #git
15:09 c0c0 joined #git
15:09 pur3eval joined #git
15:10 marcogmonteiro joined #git
15:10 PCatinean joined #git
15:12 GodGinrai joined #git
15:14 altendky joined #git
15:16 roidal iam asking because iam wondering whats the maximum amount of objects a repository can contain
15:17 altendky my coworker/boss is switching from tortoisegit on windows to linux and is having trouble finding dag visualization which focuses on branches/merges and collapses intermediate commits.  any suggestions?  https://tortoisegit.org/docs/tortoisegit/tgit-dug-revgraph.html  he's working to get to a sane model but for now the branches are, well, not sane
15:19 _ikke_ I don't know anything similar to that
15:21 shgysk8zer0_ joined #git
15:21 CheckDavid joined #git
15:22 DaveTaboola joined #git
15:23 BSab joined #git
15:24 jast the structure of that looks like: git log --simplify-by-decoration --graph --oneline --pretty=%d
15:24 mischat_ joined #git
15:26 altendky jast: i added --all and that looks like you're right.  perhaps i can get somewhere with that
15:26 jast gitk might support these flags, too, if you want something slightly more graphical
15:26 altendky jast: that was my next step, thanks  :]
15:26 jast not sure about --oneline and --pretty in particular
15:27 daivyk joined #git
15:27 _ikke_ jast: I was thinking about that too, but this is a more 'exploded' view
15:27 nettoweb joined #git
15:28 altendky jast: git-dag --all --simplify-by-decoration  (i'll see what they think of this) https://usercontent.irccloud-cdn.com/file/ZAlidk87/
15:28 marcogmonteiro joined #git
15:31 Vampire0 joined #git
15:31 durham joined #git
15:32 overlord_tm joined #git
15:34 aLeSD tobiasvl, dunno ... but I missed up with rebasing ... I just run a reset
15:34 aLeSD all good now
15:36 mknooihuisen joined #git
15:39 RxMcDonald joined #git
15:40 RxMcDonald how do you pass strings to git grep? http://pastebin.com/raw/TetGN1DB it's changing to $DIR instead of the contents of the variable lmao
15:40 babyswizz I am trying to push a local repo with a new user on git but I get remote: Permission to MYUSERNAME/project.git denied to OLD-USER-NAME error
15:40 babyswizz any idea what is wrong?
15:40 aLeSD May I push master branch meanwhile I am working on a local branhc ?
15:41 m4sk1n joined #git
15:41 Guest9 joined #git
15:42 aLeSD the prob is that push is tacking a lot of time and I 'd like to work in the meantime
15:42 mgedmin RxMcDonald: it's a shell quoting question, nothing git-specific actually: $DIR is expanded inside "-quotes, but not inside '-quotes
15:43 cdg joined #git
15:44 RxMcDonald mgedmin: nice, it works, thanks
15:44 Snugglebash joined #git
15:46 ikelso joined #git
15:46 tang^ joined #git
15:48 cdg joined #git
15:48 JamesBaxter joined #git
15:49 nckpz joined #git
15:51 Snugglebash joined #git
15:54 Vampire0_ joined #git
15:57 str joined #git
15:59 crose joined #git
16:00 zivester joined #git
16:03 Shashikant86 joined #git
16:05 ocbtec joined #git
16:10 pur3eval joined #git
16:10 dreiss joined #git
16:12 hobodave joined #git
16:12 b100s joined #git
16:12 b100s hi2all
16:13 b100s how can i see changes staged and not by one comand?
16:13 GodGinrai b100s: what do you mean by "not by one command"?
16:13 ToxicFrog GodGinrai: I think the binding is (changes, (staged and not), by one command)
16:13 ToxicFrog b100s: "git status
16:13 GodGinrai ah
16:14 GodGinrai in that case, yea, git status
16:14 ToxicFrog b100s: "git status" will list them by file. "git diff HEAD" will show you the diff in your working tree (i.e. all staged and unstaged changes) relative to the latest commit.
16:15 ToxicFrog If you want the diffs, but want separate diffs for staged and unstaged changes, I think you need two commands for that.
16:15 b100s oh, my english
16:15 b100s sorry; i mean by one command
16:16 mathematic-alpha joined #git
16:16 metachr0n joined #git
16:16 Snugglebash joined #git
16:19 publio joined #git
16:20 theelous3 joined #git
16:21 thierryp joined #git
16:22 brent__ joined #git
16:22 borkr joined #git
16:22 theelous3 in my project's directory, I've done "git add.", "git commit -m "blah", and "git push origin master". Judging by the message returned, it looks like everything went normally, however there are no changes to my repo online
16:22 menip joined #git
16:22 theelous3 trying again tells me that there's nothing to commit
16:22 selckin !bare
16:22 gitinfo A bare repository is used to push/fetch (useful for running a git server), and contains only the contents of .git/ from a "normal" repo. Read more: http://bare-vs-nonbare.gitrecipes.de/
16:22 selckin !deploy
16:22 gitinfo Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html
16:24 theelous3 well, that is confusing
16:25 theelous3 why has this process verbatim been working fine until now?
16:25 synthroid joined #git
16:25 vuoto joined #git
16:26 theelous3 I can see the file changes etc. in my repo's commits, but the files themselves haven't been altered.
16:27 theelous3 I don't know enough about git to understand the !links you gave me. If there's something more basic I'd love to see it :)
16:29 eihabarabia joined #git
16:31 eihabara_ joined #git
16:31 Hudu joined #git
16:32 sentriz joined #git
16:33 jophish I'm trying to resurrect a three year old branch of some c++ code and it's proving very difficult, has anyone got any tips for this kind of job?
16:33 theelous3 nvm, just github being really fucking slow
16:34 GodGinrai jophish: you will need a lot of patience
16:35 duderonomy joined #git
16:36 thiago joined #git
16:37 tlaxkit joined #git
16:39 jophish GodGinrai: hopefully it'll be worth it
16:40 tvl joined #git
16:41 eroussel joined #git
16:41 malice` joined #git
16:43 malice` Hi, I want to revert file to certain version in history. I tried "git checkout hash file" and it theoretically works
16:43 malice` but if file was not existing back then, it won't work. How can I fix this?
16:43 malice` I mean, I want it to become an empty file if it did not exist.
16:43 _ikke_ malice`: use git log <file> to find the commit that has the version you want
16:43 _ikke_ hmm
16:44 _ikke_ I don't think git supports that
16:44 ShalokShalom joined #git
16:44 mischat joined #git
16:44 ojacobson malice`: on a file-by-file basis, you pretty much have to do it the hard way
16:44 diogenese joined #git
16:44 robotroll joined #git
16:44 ojacobson on a whole-tree basis, `git rm -rf :/ && git checkout <commit> -- :/` will re-stage and restore the entire tree as it was in <commit>
16:44 malice` I am asking because I have a list of files I want to commit this operation on
16:45 malice` ojacobson: I appreciate the " :/ " at the end
16:45 ojacobson if ! git checkout <commit> -- path/to/file; then rm path/to/file; fi
16:45 _ikke_ ojacobson: they want an empty file, not remove it
16:46 ojacobson I have faith that malice` can figure out how to modify the script :)
16:47 madewokherd joined #git
16:47 tyreld joined #git
16:47 _ikke_ git show <commit>:<file> 2>/dev/null ><file>
16:47 ojacobson Nice.
16:48 malice` ojacobson: I did, thanks
16:49 malice` _ikke_: nice!
16:54 sitaram /away/away
16:54 sitaram oops
16:54 sitaram sorry
16:54 _ikke_ ha
16:55 ShalokShalom joined #git
16:59 hobodave joined #git
17:00 ShalokShalom joined #git
17:00 mehola joined #git
17:00 theelous3 left #git
17:09 ac3takwas joined #git
17:09 Hink joined #git
17:11 DieguezZ joined #git
17:11 umbSublime joined #git
17:12 sgen joined #git
17:12 NullableTruth joined #git
17:12 vuoto joined #git
17:12 rsmarples joined #git
17:12 rsmarples hi!
17:12 gitinfo rsmarples: 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.
17:13 rsmarples I'm struggling to setup gitweb and cgit to show my bare bones repos. both say there are no commits, but git log says otherwise.
17:14 rsmarples using 2.11.1 for gitweb and 1.1 for cgit
17:14 rsmarples https://roy.marples.name/gitweb?p=test.git;a=summary
17:15 Snugglebash joined #git
17:15 grawity can your web server actually read the repository's files?
17:15 grawity that is, permissions and stuff
17:15 rsmarples http://pastebin.com/J39XRGFV
17:15 rsmarples yes
17:16 rsmarples all files are world readible
17:17 ojacobson Are all of the directories leading to your repos traversable by the web server's user or group?
17:17 grawity but are all the directories at least world-executable? /var/git itself included
17:17 ojacobson /, /var, and /var/git, in fact :)
17:18 grawity generally I'd probably dig out `strace` and see where it's *trying* to look, anyway
17:18 grawity because if neither of the two cgi scripts can access it, it's probably something external
17:18 _ikke_ Don't you also need to mark the repo for it to show in those kidn of projects?
17:19 _ikke_ like with git daemon
17:19 rsmarples have to be kdump this is a netbsd box
17:19 ojacobson _ikke_: that generally causes the whole repo to be omitted
17:19 ojacobson not just the history :)
17:19 _noblegas joined #git
17:19 sangy is there a way to fetch an arbitrary commit object?
17:19 grawity ktrace/ktruss, I guess
17:20 grawity sangy: usually no
17:20 sangy grawity: that's what I was afraid of. Is there a way to do it artificially? (e.g,. put a ref pointing to it and fetching)
17:20 sangy or should I instead use the github api to query information about it
17:20 grawity sangy: that's how regular cloning works, isn't it
17:21 jstimm joined #git
17:21 sangy oh, of course lol. Thanks
17:21 rsmarples you can clone the repo ... git daemon runs as as the www user git://roy.marples.name/var/git/test.git
17:21 rsmarples so that means the permissions are fine
17:21 mda1_ joined #git
17:22 xall joined #git
17:22 grawity from the "Connection refused" I wouldn't say it runs
17:23 jagob joined #git
17:25 benjwadams joined #git
17:26 invisbl_ joined #git
17:26 benjwadams My terminal appeared frozen during a git push, so i ctrl-c'ed.  My repo is not showing a the new commit, but git push is claiming everything is up to date
17:26 benjwadams any way to redo a push?
17:26 rsmarples grawity: ah, one moment. opening firewall
17:27 ChristopheT joined #git
17:27 Boniche- left #git
17:27 invisbl__ joined #git
17:27 ojacobson rsmarples: I would be very hesitant to conclude anything about the http user from the success or failure of a clone using a service running as a git-daemon user
17:27 Boniche joined #git
17:27 ojacobson unless, on your server, git:// and http:// services run as the same user
17:28 jstimm joined #git
17:28 rsmarples ojacobson: same suer
17:28 rsmarples user
17:28 rsmarples grawity: try again plz
17:29 grawity hmm
17:29 grawity well, if it's NetBSD, guess you can rule out SELinux/AppArmor being an arse
17:31 tmk joined #git
17:31 sangy second random question. What'd be the easiest way to find the common antecessor of two brancheS?
17:31 sangy I'm looking at the git log manpage but I don't know if there's something easier
17:31 benjwadams never mind, things cleared up
17:32 sangy nvm found it, thanks!
17:33 roidal does the fetch/pull protocol allow to transmit more than one pack-file per request?
17:33 nettoweb joined #git
17:33 jknetl joined #git
17:33 j7k6 joined #git
17:34 roidal iam asking because iam wondering whats the maximum amount of objects a repository can contain
17:34 ojacobson roidal: No, but it does support computing a pack on demand
17:34 ojacobson The formal maximum is 2**160 objects
17:34 ojacobson any real repository will reach implementation limits (available disk space, etc) well before that point though
17:35 roidal why 2**160?
17:35 ojacobson Size of a SHA-1 hash in bits.
17:35 ojacobson Each distinct hash must indentify at most one object
17:35 roidal i was thinking about the limit of objects withing one pack file
17:35 roidal which is 4G?
17:35 roidal when fetch/pull can only transmit one file, so this is the limit?
17:36 roidal am i wrong?
17:38 mathematic-alpha left #git
17:38 roidal ojacobson: maybe i was unclear. a clone need to transfere all files within one pack-File?
17:39 roidal a pack-File has a 4-Byte object-count value, which limits it to 4G files?
17:40 cdg_ joined #git
17:40 rsmarples what git commands does gitweb do anyway? i struggle to read perl
17:40 ojacobson roidal: the (current) fetch protocol negotiates the objects to pack, then sends a single pack stream https://github.com/git/git/blob/master/Documentation/technical/pack-protocol.txt
17:41 pur3eval joined #git
17:41 canton7 roidal, I think they mean 4 billion files, not 4GB of data
17:41 ojacobson roidal: I strongly suspect most repos will hit limits due to network reliability before they hit limits due to pack file size, but you might be onto something
17:41 roidal canton7: yes, i meant 4 billion files
17:41 roidal ähm
17:42 roidal 4 billion objects
17:42 canton7 yeah, objects
17:42 matsaman joined #git
17:43 cdown joined #git
17:43 roidal and iam wondering if the fetch-protocol's "pack stream" is exactly the same like the pack-file
17:43 roidal in that case a fetch/pull might only transfere up to 4 billion objects?
17:43 canton7 yeah, but you'll likely run out of memory trying to repack it way before then :P
17:44 canton7 reading around, the practical limits around in the millions of commits and files
17:44 finalbeta joined #git
17:45 roidal i would say a 200GB pack-file is able to contain this amount of objects (assuming small files...about 50 bytes)
17:45 roidal i was less thinking on a real case scenario, more technically...
17:46 roidal ...if my understanding of this 4 billion object limit is correct, or if i have anything overseen
17:46 canton7 don't forget the object headers
17:46 canton7 ... but then, don't forget that objects can be deltafied
17:46 roidal sure
17:46 jccn joined #git
17:51 invisbl joined #git
17:51 rorro joined #git
17:52 miczac\away joined #git
17:52 e14 joined #git
17:53 invisbl joined #git
17:53 mathematic-alpha joined #git
17:53 gopar joined #git
17:54 Darren__ joined #git
17:59 Naan joined #git
18:00 mmlb joined #git
18:00 ac3takwas joined #git
18:01 nickabbey joined #git
18:01 ihradek joined #git
18:02 dreiss joined #git
18:05 amagawdd_ joined #git
18:08 jstimm joined #git
18:09 chardan joined #git
18:10 jnavila joined #git
18:13 pandeiro joined #git
18:13 hahuang65 joined #git
18:14 mknooihuisen joined #git
18:15 ChristopheT left #git
18:17 eroussel joined #git
18:17 j7k6_ joined #git
18:18 synthroi_ joined #git
18:19 joshszep joined #git
18:20 rsmarples solved the issue, noob mistake
18:20 Murii_ joined #git
18:20 rsmarples compiled git for a slightly newer libc
18:26 evie_hammond joined #git
18:26 MacWinner joined #git
18:27 F0rTh3J3st joined #git
18:30 spacelord_ joined #git
18:30 matsaman rsmarples: what was the issue?
18:30 rsmarples matsaman: my buildbox had a slightly newer libc than the running host
18:30 rsmarples caused git itself to segfault
18:30 matsaman oh, that's a neat one
18:30 rsmarples but only for calls generated by gitweb
18:30 matsaman aha
18:30 matsaman that makes trivially more sense
18:31 * rsmarples wonders what changed in the one day of libc
18:31 mknooihuisen left #git
18:32 mischat joined #git
18:32 mischat joined #git
18:35 duderonomy joined #git
18:36 Snugglebash joined #git
18:37 lmat joined #git
18:37 Nokurn joined #git
18:37 KevinDF joined #git
18:38 Cabanossi joined #git
18:39 mischat joined #git
18:39 matoro joined #git
18:40 dreiss joined #git
18:40 synthroid joined #git
18:40 heroux joined #git
18:40 Impaloo joined #git
18:43 ignacio left #git
18:44 spacelord_ joined #git
18:45 GodGinra1 joined #git
18:45 a-l-e joined #git
18:46 JanC_ joined #git
18:46 vdv joined #git
18:49 notebox joined #git
18:50 GodGinrai joined #git
18:53 raynold[here] joined #git
18:54 raynold[here] left #git
18:54 mknooihuisen joined #git
18:54 raynold[here] joined #git
18:55 raedah joined #git
18:56 jstimm joined #git
18:56 cdg joined #git
18:57 Tobbi joined #git
18:58 spacelord_ joined #git
18:59 jimi_ joined #git
19:01 jaafar joined #git
19:02 GodGinrai joined #git
19:03 jstimm joined #git
19:04 canuk joined #git
19:04 Nokurn joined #git
19:05 MineCoins joined #git
19:07 finalbeta joined #git
19:08 Orphis joined #git
19:09 sathed joined #git
19:10 amagawdd joined #git
19:10 ozmage joined #git
19:11 Vampire0_ joined #git
19:11 Sasazuka joined #git
19:11 pur3eval joined #git
19:13 irqq joined #git
19:13 matoro joined #git
19:14 jimsio joined #git
19:15 dunk joined #git
19:15 moop joined #git
19:17 danimal2 joined #git
19:18 [diecast] joined #git
19:19 HardlySeen left #git
19:19 nostrora joined #git
19:20 alex1a joined #git
19:21 basiclaser joined #git
19:21 johtso joined #git
19:21 fcrick joined #git
19:22 adambrenecki joined #git
19:22 kwmiebach___ joined #git
19:22 mroth joined #git
19:22 shennyg joined #git
19:22 manuelschneid3r joined #git
19:22 XDS2010 joined #git
19:22 lebster joined #git
19:22 stealthii joined #git
19:23 bcardarella joined #git
19:23 chachasmooth joined #git
19:23 zifnab joined #git
19:23 ec joined #git
19:23 GameGamer43 joined #git
19:23 sr___ joined #git
19:23 huhlig joined #git
19:24 bbhoss joined #git
19:25 CheckDavid joined #git
19:25 devhost joined #git
19:25 pizzaops joined #git
19:25 mischat joined #git
19:25 mostlybadfly joined #git
19:26 addshore joined #git
19:26 gyre007 joined #git
19:27 vcoinminer joined #git
19:28 ggherdov joined #git
19:29 Jackneill_ joined #git
19:30 manveru joined #git
19:30 ome joined #git
19:31 ionelmc joined #git
19:31 DuncanT joined #git
19:31 rfv joined #git
19:31 Tee_Pee joined #git
19:32 nikivi joined #git
19:32 angelsl joined #git
19:33 Whisket joined #git
19:34 beckyconning___ joined #git
19:34 cstrahan joined #git
19:34 daxroc joined #git
19:35 epic__ joined #git
19:36 phtes joined #git
19:36 navidr joined #git
19:36 buckowski joined #git
19:36 Lowell joined #git
19:37 Tobbi joined #git
19:37 pfg joined #git
19:37 eshlox joined #git
19:39 cloudcell_ joined #git
19:40 Nilesh_ joined #git
19:41 j7k6__ joined #git
19:41 basiclaser joined #git
19:42 _noblegas joined #git
19:42 marc_v92 joined #git
19:42 chardan joined #git
19:44 contracode joined #git
19:45 CheckDavid joined #git
19:45 dpower joined #git
19:47 sostenuto joined #git
19:48 dviola joined #git
19:49 kfc joined #git
19:50 nezZario joined #git
19:50 martin_g joined #git
19:50 ome joined #git
19:51 patarr joined #git
19:51 patarr Hello. Let's say I made local changes while I was in a detached HEAD state. Can I get back up to HEAD without throwing away my local changes?
19:51 DuncanT joined #git
19:51 ojacobson !reattach
19:51 gitinfo Letters refer to !detached. (a) and (b): 'git checkout branchname' to continue working on another branch, or 'git checkout -b branchname' to start a new one here; (c) git am --continue; (d) git rebase --continue
19:52 worstadmin joined #git
19:52 ojacobson probably the "start a new one here" case
19:52 barq joined #git
19:52 patarr I don't want to start a new branch
19:52 ojacobson a detached head is a branch with no name anyways, "keeping" it is a matter of giving it a name
19:52 ToxicFrog patarr: by "back to HEAD" did you mean "back to an existing branch"? "HEAD" is just "whatever commit is currently checked out"
19:53 ToxicFrog If so, you might just be able to "git checkout <branch>" and if safe git will float your uncommitted changes over to the branch.
19:53 graingert__ joined #git
19:53 ToxicFrog If that's not safe, git will refuse to do it and you will probably need to "git stash", "git checkout" and then "git stash pop".
19:53 Travis-42 joined #git
19:53 patarr ToxicFrog: that's what I'm thinking, but I was just worried that when I do the checkout it will blow away my local uncommitted changes
19:53 ToxicFrog patarr: if it would destroy uncommitted changes it will refuse to do the checkout.
19:53 ojacobson nope, !float applies
19:53 gitinfo If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
19:54 patarr Let's just hope git has my best interests at heart and looks out for me here. Thanks a bunch ToxicFrog and ojacobson
19:54 marc_v92 joined #git
19:54 aendrew joined #git
19:54 ojacobson o7
19:54 atomi joined #git
19:54 buckowski joined #git
19:54 wenzowski joined #git
19:55 dmj` joined #git
19:56 cmosguy joined #git
19:56 ulkesh joined #git
19:57 m4sk1n joined #git
19:58 BackEndCoder joined #git
19:58 _ikke_ patarr: new objects are not removed for 14 days. The reflog even holds on to objects for 90 days
19:58 _ikke_ So I would say, git got you covered
19:58 kutenai joined #git
19:59 cdown joined #git
19:59 stevenxl joined #git
19:59 notbenh_ joined #git
19:59 chigang joined #git
20:00 kireevco joined #git
20:00 sharon_so joined #git
20:00 qqx As long as you've `git add`ed the content anyway.
20:00 [mgv] joined #git
20:00 qqx Although it can be difficult to find the appropriate blob for uncommitted changes.
20:01 atomi joined #git
20:01 etcinit joined #git
20:02 sfan786_ joined #git
20:02 flupke joined #git
20:02 tharuin joined #git
20:03 _ikke_ git fsck --lost-found
20:03 codedmart joined #git
20:03 durham joined #git
20:03 boxrick1 joined #git
20:03 Nokurn joined #git
20:03 iota_pi joined #git
20:05 llamapixel joined #git
20:05 aleb joined #git
20:05 sebhoss joined #git
20:05 timwis joined #git
20:06 _joes__ joined #git
20:07 altendky joined #git
20:07 cttttt joined #git
20:07 stoopkid joined #git
20:08 discopatrick joined #git
20:08 unholycrab joined #git
20:09 al-damiri joined #git
20:09 aidalgol joined #git
20:09 parliament joined #git
20:10 solenodic joined #git
20:10 t-richards joined #git
20:11 manuelschneid3r joined #git
20:12 jleclanche joined #git
20:12 lolmac joined #git
20:13 milkshoes joined #git
20:14 ozmage joined #git
20:14 hussam joined #git
20:14 Ferris__ joined #git
20:16 LiftLeft joined #git
20:17 _ADN_ joined #git
20:17 tvw joined #git
20:18 davisonio joined #git
20:18 gajus joined #git
20:18 matsaman joined #git
20:18 belak joined #git
20:19 dtype joined #git
20:20 m0viefreak joined #git
20:20 valize joined #git
20:23 Levex joined #git
20:24 hvxgr joined #git
20:24 raynold[here] joined #git
20:24 GodGinrai joined #git
20:24 vuoto joined #git
20:26 GodGinrai joined #git
20:28 gugah joined #git
20:28 Sound joined #git
20:28 gugah joined #git
20:30 Nokurn joined #git
20:32 rivarun joined #git
20:32 jstimm joined #git
20:33 gigq joined #git
20:33 ertes joined #git
20:33 raynold sup
20:34 Rodya_ joined #git
20:35 dave0x6d joined #git
20:38 nickabbey joined #git
20:39 ankit01ojha joined #git
20:41 Murr joined #git
20:42 krainboltgreene joined #git
20:42 krainboltgreene Is it feasible to create a branch and commit in a single action without moving from your current branch?
20:42 grawity depends on what you consider a single action
20:43 krainboltgreene I'm on `master` and I have several modified/untracked files. Each of them should be their own branch/PR. It would take me quite a lot of time to create a branch for each manually, then commit manually.
20:43 grawity yes it's doable without changing HEAD or the worktree, but is a bit manual
20:43 krainboltgreene I basically want `git commit -m "x" -b y`
20:43 grawity (sidenote: you can have multiple worktrees now)
20:43 krainboltgreene Oh really?
20:44 grawity git branch foo $(git commit-tree -p HEAD -m message $(git write-tree))
20:44 grawity man git worktree
20:44 gitinfo the git-worktree manpage is available at http://jk.gs/git-worktree.html
20:44 Gloomy joined #git
20:44 ozmage joined #git
20:44 grawity hmm no, commit-tree wants it as stdin
20:44 Tobbi joined #git
20:47 krainboltgreene I guess I can do it as a bash script.
20:48 krainboltgreene Initial thoughts: https://gist.github.com/krainboltgreene/2961722a769a2e411bca36d83234da41
20:55 Vampire0_ joined #git
20:55 _ADN_ joined #git
20:57 shgysk8zer0__ joined #git
20:57 nickabbey joined #git
21:00 pur3eval joined #git
21:00 [ADN] joined #git
21:01 Vampire0 joined #git
21:01 mehola joined #git
21:01 kyan joined #git
21:03 [ADN] joined #git
21:06 p4trick joined #git
21:09 gugah joined #git
21:12 prg3 joined #git
21:13 dviola joined #git
21:14 diego3 joined #git
21:14 matsaman joined #git
21:14 diego3 joined #git
21:15 matsaman joined #git
21:15 askb joined #git
21:16 gigq joined #git
21:18 cdg joined #git
21:18 jetpack_joe joined #git
21:19 cdg_ joined #git
21:20 dsdeiz joined #git
21:21 Xeon06 joined #git
21:21 jetpack_joe joined #git
21:21 krainboltgreene Does anyone know of a "better" `git add -i`?
21:21 krainboltgreene It seems...complex.
21:22 Xeon06 Hey guys. I merged a branch using git merge and there was no merge commit created. Instead all the individual commits from the branch I was merging from moved to the branch I was merging to. What happened?
21:22 Atm0spher1c joined #git
21:22 matsaman krainboltgreene: why interactively?
21:22 hashpuppy joined #git
21:22 krainboltgreene matsaman: I want to add only certain bits of a working tree.
21:22 matsaman Xeon06: sounds like you rebased; what'd you actually run?
21:23 matsaman krainboltgreene: why not just git add the files in question
21:23 krainboltgreene matsaman: Because I only want certain hunks of a working tree.
21:23 Xeon06 matsaman: I know right? But I checked my command history and it is:  643  git checkout master   644  git merge css-rewrite   645  git push
21:23 matsaman well that's complicated because you've made it complicated
21:23 sonOfRa Xeon06: sounds like a fast-forward merge
21:24 krainboltgreene matsaman: I didn't make git add -i.
21:24 krainboltgreene Github's GUI makes it pretty easy.
21:24 krainboltgreene Same with Gitx.
21:24 matsaman krainboltgreene: no, and anything requiring a -foo in git is complicated; but you _did_ make your complicated situation
21:24 sonOfRa if you want merge commits every time, even if a fast-forward merge is possible, use git merge --no-ff
21:24 krainboltgreene Okay, well it's easy with other non-cli tools, soooo it must not be that complciated a situation.
21:24 Xeon06 sonOfRa: Ohhh I see, that's probably what happened
21:25 krainboltgreene matsaman: But like, feel free to blame me for git add -i being awkward.
21:25 Xeon06 Okay, my actual question is: How can I checkout master to right before merging, if there is no merge commit?
21:25 matsaman nothing is complicated with a willingness to store infinite redundant information in your head =)
21:25 matsaman krainboltgreene: everything with a - is awkward in git, I recommend avoiding it
21:25 gigq joined #git
21:25 krainboltgreene What in the world does that even mean in this context?
21:25 matsaman I mean if you are happy to remember minutia, nothing is complicated =)
21:26 durham joined #git
21:26 sonOfRa You can also toggle this in gitconfig
21:27 sonOfRa git config merge.ff true/false toggles whether you want to allow this
21:27 matsaman that way you can train yourself that 'git merge' isn't what it is by default =)
21:27 Xeon06 sonOfRa: I usually wouldn't mind but my problem now is that I wanna checkout to repository to right before that merge, and I don't know how to do that anymore without a single commit
21:27 frobnic joined #git
21:28 krainboltgreene Ah-ha, `git gui` works great!
21:28 Levex joined #git
21:28 regedit joined #git
21:29 sonOfRa Xeon06: yes, this is a problem I run into frequently as well, which is why I disabled fast-forward merges for a lot of my repositories
21:29 sonOfRa If you haven't pushed the merge to a remote yet, you can undo this
21:29 matsaman that git merge has ff & no-ff is bad design, IMO =/
21:29 Xeon06 sonOfRa: I have unfortunately :(
21:29 matsaman rather, that it does one or the other based on some magical calculation
21:29 sonOfRa If you've pushed it, you can still undo it, but you will be rewriting public history
21:29 sonOfRa matsaman: it's not exactly magical
21:29 matsaman it should always do one or the other
21:29 matsaman ^or it's magic
21:30 sonOfRa It's simply "is this commit an ancestor of the branch you're merging"
21:30 Xeon06 sonOfRa: I only need to temporarily check out what it was before the merge, can I do that without rewriting history?
21:30 matsaman why would I care about that
21:30 jeffreylevesque joined #git
21:30 matsaman a tool should do a thing reliably
21:30 sonOfRa Xeon06: well, locate the commit with git log, and checkout that
21:30 Xeon06 sonOfRa: But master and my feature branch's commit are intermixed now
21:31 sonOfRa Xeon06: are you sure you didn't rebase? Because "intermixed" sounds like rebase, not ff-merge
21:31 Xeon06 100% sure
21:31 Xeon06 643  git checkout master   644  git merge css-rewrite   645  git push
21:31 durham joined #git
21:31 sonOfRa matsaman: ff is a very reasonable feature, especially for pulling
21:31 matsaman especially, or only?
21:31 mischat joined #git
21:32 matsaman if the way it were utilized or not were reasonable, there wouldn't be people coming in here confused about it
21:32 sonOfRa especially. There's nothing really wrong with fast-forwarding a merge that's simply an ancestor of the current HEAD
21:32 matsaman no, there's nothing wrong with anything if you know everything ahead of time and have infinite memory =P
21:34 Levex joined #git
21:34 Balliad joined #git
21:34 Xeon06 Wouldn't a FF merge cause an intermixing of the commits?
21:35 matsaman a lot of people all over the place complain about fast forward
21:36 nowhereman joined #git
21:38 cdg joined #git
21:40 sonOfRa matsaman: have you ever tried bisecting a master branch that consists of a bunch of no-ff merge commits?
21:40 sonOfRa Pain beyond measure.
21:41 sonOfRa git blame and git bisect behave really weirdly in the presence of a lot of merge commits that could have been fast forwards
21:41 Xunie joined #git
21:41 AndresInSpace joined #git
21:41 nettoweb joined #git
21:43 matsaman why's that
21:43 cdg_ joined #git
21:45 sonOfRa Simply because you lose information in a merge commit that a fast-forward merge retains
21:45 sonOfRa Let's say, your feature branch consists of 10 commits, and you ff-merge into master. Then you can, when bisecting, or blaming, select each of those 10 commits to figure out which one introduced a regression
21:46 minn joined #git
21:46 matsaman rather than it being just one commit
21:46 sonOfRa If you --no-ff merge it, you have one big commit
21:46 sonOfRa And bisect becomes useless. You know "this merge introduced it", but you don't know which part of the merge did it.
21:46 matsaman I can see how the behavior of particular contributors could make that annoying
21:47 matsaman again, though, primarily it's that it's not applied the same way in all situations that bothers me
21:47 sonOfRa So, it depends very very much on the workflow you have decided on. And for most workflows, the default behaviour - fast-forwarding by default - is a good default, because it retains these capabilities
21:48 icey left #git
21:50 AndresInSpace OH thats why
21:50 j7k6 joined #git
21:50 AndresInSpace making me login..
21:51 ojacobson sonOfRa: git-bisect will explore both arms of a non-fast-forward merge
21:51 ojacobson If you have a test case showing otherwise, please pass it to the list - that would be a bug
21:51 ojacobson (for "both" read "all" if it's an octopus merge)
21:51 matsaman that's a human bug anyway, IMO
21:51 matsaman ff being applied or not because of a calculation is a proper software bug =) (feature? =P)
21:52 AndresInSpace left #git
21:52 ojacobson Personally, I agree - my git config disables ff merges (I have aliases that enable it for specific commands)
21:52 ojacobson The "create a merge sometimes but not always" default has always struck me as weird
21:53 matsaman hear hear
21:53 matsaman and let's stone those who disagree! =P
21:53 mostlybadfly joined #git
21:53 vlb joined #git
21:53 vlb left #git
21:53 vlb joined #git
21:53 vlb left #git
21:53 matsaman </explicitcloseofjokeforthesetroublingtrumpdays>
21:53 ojacobson I think I'm misunderstanding sonOfRa's point, because as far as I know merge commits retain a strict superset of the data points a fast-forward merge keeps
21:53 vlb joined #git
21:53 ojacobson and I interpret their remarks as saying that merge commits throw away some info
21:53 matsaman vs --squash?
21:53 ojacobson Oh vs --squash, fine
21:54 ojacobson --squash throws out all sorts of shit, that's where my misunderstanding was
21:54 matsaman no I don't think he said that
21:54 ojacobson I missed context :)
21:54 matsaman but he might have meant it
21:54 * matsaman shrugs
21:54 * ojacobson shrugs
21:54 * matsaman shrugs the best
21:54 ojacobson I know sonOfRa from a couple of other channels and e's always been reasonably on point, so I assume I missed something there
21:54 Vampire0_ joined #git
21:55 vlb left #git
21:55 matsaman his point makes sense to me, whether it's the case or not
21:55 matsaman it just has no bearing, IMO, on ff being applied magically
21:55 matsaman it's more to do with how merges work out or don't after they happen
21:55 * matsaman separates all the things
21:56 ozmage joined #git
21:56 vlb joined #git
21:57 vlb hi
21:57 gitinfo vlb: 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.
21:58 pietv joined #git
21:58 sonOfRa I may simply be wrong on the behaviours of bisect and blame here.
22:00 ayogi joined #git
22:00 UTAN_dev joined #git
22:01 pur3eval joined #git
22:01 vlb I'm trying to do merge a feature branch and selectively take hunks from it. Any ideas how I can do this? If I do a "git merge origin/feature ; git reset ; git add -p" I then the feature branch does not appear to be merged in the history.
22:02 vlb it seems that the reset step actually remove the fact that I am merging
22:03 vlb solutins online does not appear to be really merging the feature branch as well (http://stackoverflow.com/questions/449541/how-do-you-merge-selective-files-with-git-merge)
22:03 mischat joined #git
22:04 ToxicFrog vlb: "git merge --no-commit" and then edit the staged changes before committing?
22:05 vlb ToxicFrog: I can do it with git merge --no-commit --no-ff then with git add -i
22:05 vlb but I was wondering if I could avoid the git add -i
22:05 vlb ie a non interactive way of doing it
22:06 ToxicFrog Well, at some point you need to tell it which hunks you want to merge and which you want to discard
22:06 ToxicFrog You can do that with your editor, or add -i/-p
22:06 ToxicFrog Or if you want to discard entire files, checkout
22:06 ToxicFrog But you have to do it somehow
22:06 vlb hmm true
22:06 vlb the problem is that by default all the hunks are already added
22:07 vlb i can reset all of them with git add -i
22:07 vlb but i was hoping to find a "nicer" way
22:07 vlb ie, I'm surprised that there are no command equivalent to git add -i , then select revert
22:08 blackwind_123 joined #git
22:08 leprechau joined #git
22:09 Levex joined #git
22:11 cdg joined #git
22:13 cdg__ joined #git
22:14 ToxicFrog git reset; git add -p?
22:14 ToxicFrog The commit parentage information should be preserved
22:14 cqi joined #git
22:14 ToxicFrog Alternately, git merge --no-ff, then git reset HEAD^, then git add -p, git commit --amend
22:17 vlb ToxicFrog: no, it seem sthat the commit parentage is gone with the reset
22:17 vlb lemme try with HEAD^
22:18 vlb doing a "git status" after the reset clearly show that we're not in a merge situation anymore btw
22:21 cdg joined #git
22:22 dmin7b5 joined #git
22:23 chardan joined #git
22:23 Darren__ joined #git
22:23 Balliad joined #git
22:24 mischat joined #git
22:24 svm_invictvs joined #git
22:24 nickabbe_ joined #git
22:24 Pappa joined #git
22:25 nixjdm_ joined #git
22:25 peterpp joined #git
22:25 cdg_ joined #git
22:26 Rodya_ joined #git
22:30 peterpp Hi, I'm using a Bitbucket repo with some team members, and just know I ran into a situation I can't explain: some hours ago I made a commit in which I updated a single file, and I can see the commit on Bitbucket and in in my local log, but when I do `git log -p -- <thefile>` then this commit isn't listed, and the changes made by the commit are also gone
22:30 Darcidride joined #git
22:30 peterpp even though my commit hasn't been reverted
22:33 a3Dman joined #git
22:34 Balliad joined #git
22:35 jimi_ joined #git
22:36 j7k6_ joined #git
22:36 Gloomy joined #git
22:37 kadoban peterpp: Check which branch you're on, that it matches what you're looking at on bitbucket, and check that you're in the place in the working directory that you think you are.
22:38 peterpp thanks for responding, actually we have only a single branch
22:39 gugah joined #git
22:39 peterpp git-wise this is an extremely simple and straightforward project
22:42 ocbtec joined #git
22:43 _ADN_ joined #git
22:43 peterpp kadoban, when I look at the file's history on bitbucket, the commit isn't there either
22:43 vlb left #git
22:44 paws_ joined #git
22:44 peterpp kadoban, I just manually restored the changes made be the commit that got 'lost', but this is very strange
22:49 solenodic joined #git
22:49 carcrash joined #git
22:50 carcrash is pull request a feature of git, or is that just a git hub thing? all my googling only talks about github
22:51 belak carcrash: a pull request is something that was invented by git hosting providers to make it easier to submit changes
22:51 carcrash belak, ok thanks.
22:51 peterpp other providers also offer this feature, like bitbucket
22:51 belak Since previously it had just been submitting patches to mailing list and having it merged by the maintainer
22:51 belak Yep, bitbucket, github, and gitlab (among others)
22:52 carcrash is there any likelihood of that being added to git, or is it considered out of scope?
22:52 kadoban man git request-pull
22:52 gitinfo the git-request-pull manpage is available at http://jk.gs/git-request-pull.html
22:52 Vinnie_win joined #git
22:52 carcrash oh, so it is a part of git now?
22:52 kadoban Pull requests have been a part of git for a long time, they're just usually informal. The tooling to accomplish them doesn't need to be very extensive.
22:53 kadoban "Hey, pull these changes: stuff"  there, a pull request.
22:53 mischat joined #git
22:53 re1 joined #git
22:55 frobnic joined #git
22:56 cdg joined #git
22:59 llamapixel pull is a combo of commands fetch and branch etc.
22:59 cdg__ joined #git
23:00 llamapixel fetch and merge
23:00 llamapixel https://longair.net/blog/2009/04/16/git-fetch-and-merge/
23:01 fmeerkoetter joined #git
23:02 endiruna joined #git
23:03 mehola joined #git
23:05 gopar joined #git
23:05 pwnz0r joined #git
23:07 joshszep joined #git
23:07 ShapeShifter499 joined #git
23:17 ikelso joined #git
23:19 gugah joined #git
23:21 mda1 joined #git
23:22 Tobbi joined #git
23:24 mischat joined #git
23:28 Rodya_ joined #git
23:32 pur3eval joined #git
23:38 theoceaniscool joined #git
23:38 ozmage joined #git
23:38 dsdeiz joined #git
23:38 dsdeiz joined #git
23:42 matoro joined #git
23:43 mda1 joined #git
23:44 nettoweb joined #git
23:47 chardan joined #git
23:49 peterpp what is wrong with my git? just now I made a commit, and bitbucket claims it had been made by one of my team members
23:50 peterpp how is this possible?
23:50 matsaman you did the push yourself?
23:50 peterpp yes
23:50 matsaman does git config -l look right?
23:51 peterpp yes
23:51 peterpp I made thousands of commits with the current configuration
23:51 peterpp all of a sudden I get this
23:52 bernardio joined #git
23:52 preaction same commit hash?
23:53 NoOova joined #git
23:53 NoOova Hi guys!
23:53 NoOova How to log commits that exists in my branch but missing in master?
23:54 cbreak git log master..branch
23:54 cbreak or git log ^master branch
23:54 cbreak see man gitrevision
23:54 cbreak see man gitrevisions
23:54 gitinfo the gitrevisions manpage is available at http://jk.gs/gitrevisions.html
23:55 NoOova thank =) i tries reversed order: HEAD..master. =)
23:55 NoOova tried
23:55 peterpp preaction, it's so odd...
23:56 ertes joined #git
23:56 preaction peterpp: if it's the same commit hash, then it's either somehow they got into the commit via Signed-Off-By or another commit data field, or i don't believe you
23:56 peterpp there are two different hashes
23:56 dtype for those moving from subversion and used to r##### tags to refer to checkins, is the new answer that the nature of git means that you're using the 7 char hashes for general reference, and the easy to reference thing is a thing of the past?
23:57 dtype I think that's the case, but don't have practical experience with it.
23:57 elijah joined #git
23:57 cdg joined #git
23:58 dviola joined #git
23:58 dmc joined #git
23:58 Vinnie_win_y joined #git
23:58 peterpp there's one hash ab4bedf which I see in bitbucket, according to bitbucket it was made by my team member and it contains exactly the changes that I wanted to make
23:58 Wulf Hello. I've got two separate repositories that share the same files but with different paths.  Repo A: /foo/bar/a,b,c  Repo B: /a,b,c. There was a commit to A and I'd like to copy/cherry-pick it to B somehow. Is there a better way than diff+patch which keeps meta information (author, date, etc.)?
23:59 preaction peterpp: then what happened was they rebased and changed the author of the commit. two different commits.
23:59 peterpp then there's a second commit with hash dc90e84 which I only see in my local log, which was also according to git made by my team member, but in addition to the changes I wanted to make it also contains a bunch of other stuff
23:59 preaction peterpp: or they just happened to make the same commit. point being, git's fine. this is correct behavior
23:59 jstimm joined #git

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