IRCloggy #git 2013-10-06

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2013-10-06

← bwreilly left00:02
← petey left00:02
→ v1n joined00:02
→ dyng joined00:03
← Cowmoo` left00:03
→ xobb joined00:04
← grilix_ left00:07
→ CKyle joined00:07
← dyng left00:07
→ grilix joined00:07
← alan_o left00:07
→ Risto- joined00:08
→ D-Chymera joined00:09
← CKyle left00:13
→ shinnya joined00:13
→ aslaen joined00:14
← King_Hual left00:16
← arand left00:16
→ thierryp_ joined00:17
← barq left00:17
← grilix left00:19
← skorgon left00:19
← thierryp left00:20
← GT4066 left00:21
→ grilix joined00:22
← CheckDavid left00:22
← Mars` left00:22
→ Mars` joined00:23
← Mars` left00:27
← RaceCondition left00:30
← skered- left00:30
→ PaulCapestany joined00:31
→ RaceCondition joined00:31
→ normanrichards joined00:31
→ skorgon joined00:33
→ skered- joined00:35
← DonRichie left00:36
→ DonRichie joined00:36
← travisstaton left00:37
→ angelsl joined00:38
→ CKyle joined00:40
← Davey left00:41
→ jfinn joined00:41
← CKyle left00:41
← redrocket left00:43
← Skrypter left00:43
← Artpicre left00:45
← ozette left00:45
← Milossh left00:45
← edran left00:45
← minim left00:47
→ milestone joined00:48
← dfanjul left00:48
→ Skrypter joined00:49
→ Milossh joined00:50
→ arjun_ joined00:50
imMute^ → imMute00:50
← cigarshark left00:53
→ minim joined00:55
→ CKyle joined00:55
→ rfoust joined00:57
← milestone left00:58
← FernandoBasso left01:01
← HerrK left01:03
← ErKa left01:03
← St_Marx left01:04
→ cagedwisdom joined01:05
← RobW_ left01:08
→ Vampire0 joined01:09
← mikepack left01:10
→ mikepack joined01:11
← Vampire0_ left01:11
→ Mars` joined01:11
→ travisstaton joined01:12
→ Davey joined01:12
← tagava|2 left01:13
← tagava left01:13
← iderik left01:14
→ mizu_no_oto joined01:15
← mikepack left01:16
← pbujok left01:16
← notjohn left01:18
← looper left01:19
← styles left01:20
→ looper joined01:21
→ Goplat joined01:23
→ nathancahill joined01:24
← ItSANgo left01:24
→ ItSANgo_ joined01:25
← oojacoboo left01:25
← Misan left01:26
← nathancahill left01:26
← frgtn left01:28
→ frgtn joined01:28
← fuzzybunny69y left01:29
← Mars` left01:29
→ Mars` joined01:30
→ oojacoboo joined01:31
← frgtn left01:33
→ dyng joined01:33
→ intripoon joined01:34
← Mars` left01:34
← RaceCondition left01:34
← intripoon_ left01:37
← mitsuhiko left01:38
← dyng left01:38
← ska-fan left01:38
→ ska-fan joined01:39
→ mitsuhiko_ joined01:40
← chrisss404 left01:44
← minim left01:46
→ m0ltar joined01:47
→ chrisss404 joined01:48
→ R1KRDO joined01:49
← looper left01:50
← NicolasTarzia left01:50
→ NicolasTarzia joined01:50
→ looper joined01:51
← oonska left01:51
← CKyle left01:51
→ threesome joined01:51
→ minim joined01:52
→ oonska joined01:52
→ tmcmahon joined01:52
← chrisss404 left01:52
← NicolasTarzia left01:53
← kexd left01:53
→ NicolasTarzia joined01:53
→ milestone joined01:54
← R1KRDO left01:55
← minim left01:56
← kpreid left01:56
← tmcmahon left01:57
← milestone left01:59
→ chrisss404 joined01:59
→ frgtn joined01:59
← threesome left01:59
→ kpreid joined02:00
← shabob left02:02
→ Prudhvi joined02:03
→ shabob joined02:04
→ Alenah joined02:04
← kwmiebach left02:04
→ spiff joined02:05
← f8l left02:06
← akitada left02:07
← skelterjohn left02:07
← frgtn left02:07
← robstanley left02:07
→ f8l joined02:07
gitinfo set mode: +v02:07
← rryoumaa left02:08
← mattp_ left02:08
← jds left02:08
→ skelterjohn joined02:08
→ robstanley joined02:08
→ jds joined02:08
← penguin42 left02:08
→ akitada joined02:08
spiff So I want to set up a repo that looks something like lib/ lib/perl/ lib/python/ lib/tcl/ etc...02:09
→ mattp_ joined02:09
→ rkpai joined02:09
spiff where contribs can work on particular languages and not have access to other languages02:10
should I use seperate repos? or use branches?02:10
pdurbin separate repos or gitolite02:10
spiff in the branches scenerio lib/ would be "master"02:11
@pdurbin - thats what I was thinking02:11
would there be any benefit to branching the languages?02:12
→ crocket joined02:12
→ mribeirodantas joined02:12
spiff eventually the whole set would need to be distributed as a package02:13
crocket http://pastebin.com/bfctSc3n is my ~/.gitconfig. However, git doesn't seem to use git-credential-gnome-keyring02:13
After I type in a password, a password is not saved.02:13
← aslaen left02:14
→ errr joined02:14
pdurbin spiff: controling access per branch is not something you can do with git out of the box, from what I understand02:14
Hello71 git doesn't *have* access control02:15
→ CKyle joined02:15
← chrisss404 left02:15
spiff thats my understanding as well, developers will only work on one language, I'll be maintaining the whole and releasing02:16
pdurbin spiff: but you don't want the python people to even see the perl code?02:16
→ chrisss404 joined02:17
spiff i suppose they could, what i'm trying to accomplish (if I go with branches) is that the lib/perl/ branch can be independently fetched/pulled with out the other languages. I dont think this is really feasible (and probably not doable), it was just a thought with this new project02:19
crocket Does anyone know how to make git-credential-gnome-keyring work?02:19
→ RobW_ joined02:20
pdurbin spiff: sure, you can have the perl people work on branch called "perl" ... and merge it in to master at some point02:21
← CKyle left02:21
→ cads joined02:21
→ Dougie1871 joined02:21
spiff pdurbin: I guess I got to think about this a little more, but it's dinner time. enough work for Saturday! :)02:22
→ diegoviola joined02:22
pdurbin crocket: sorry, I just use ssh-add (ssh-agent) before I git push or whatever02:22
spiff: good luck02:22
← spiff left02:23
← mfulz left02:23
→ bwreilly joined02:24
← rkpai left02:25
← bwreilly left02:26
← relipse left02:27
→ hanser_ joined02:27
→ relipse joined02:27
→ dyng joined02:27
← f8l left02:30
→ f8ld joined02:30
→ amunoz joined02:30
← hanser left02:30
→ frsk joined02:30
→ yhager|away joined02:32
crocket pdurbin, ssh-add?02:32
meh02:33
pdurbin crocket: ssh-add is at https://help.github.com/articles/generating-ssh-keys02:34
crocket pdurbin, /usr/share/doc/git/contrib/credential/gnome-keyring/git-credential-gnome-keyring02:34
← RobW_ left02:37
→ dreiss joined02:39
→ Mars` joined02:40
← suy left02:42
→ suy joined02:42
← yacks left02:42
← mizu_no_oto left02:44
← v1n left02:44
→ jefflyne joined02:44
← Mars` left02:46
→ |Freeman| joined02:46
← sfz left02:49
→ mikepack joined02:49
← arjun_ left02:50
← Risto- left02:50
→ rchavik joined02:50
← mobileblue left02:52
→ RobW_ joined02:53
← joki left02:54
RobW_ Is there a way to pull in all the refs / commits from a remote branch onto a local branch without checking the local branch out?02:54
→ mobileblue joined02:54
→ Shaladdle joined02:55
→ milestone joined02:55
← jason237 left02:55
RobW_ Trying to update the local branch before moving its HEAD ref to an existing working directory -- checkout without checking out.02:56
← edwardly left02:56
crocket What the hell???02:56
milki lol02:56
RobW_ has been doing strange things for a while02:56
→ jason237 joined02:57
RobW_ Ha, yeah.02:57
crocket "git push" and "git fetch" don't execute credential.helper when a remote's protocol is ssh.02:57
← m0ltar left02:57
RobW_ It's true.02:57
crocket When its protocol is https, it executes credential.helper.02:57
milki maybe02:57
i use ssh-agent for that02:57
does ssh normally execute credntial.helper?02:57
← chrisss404 left02:57
milki i probably does it if your key is id_rsa02:58
→ edwardly joined02:58
milki and if you explicitl specify a key02:58
← milestone left02:59
→ joki joined03:00
crocket milki, "git fetch" should execute credential.helper even if a remote is an SSH host.03:01
→ chrisss404 joined03:01
RobW_ … git fetch <remote> <remote branch>:<local branch> fails if the current branch is the local branch and has a dirty working directory. So I'm trying to avoid the "Refusing to fetch into current branch refs/heads/test-branch of non-bare repository" error.03:02
← dyng left03:03
RobW_ And there's only one branch in the local repo. Maybe I have to change workflow in order to avoid this.03:03
→ dyng joined03:03
crocket Does git execute credential.helper even when a remote is SSH?03:04
← donmichelangelo left03:04
→ donmichelangelo joined03:05
crocket https://help.github.com/articles/set-up-git says "The credential helper only works when you clone an HTTPS repository URL. If you use the SSH repository URL instead, SSH keys are used for authentication. This guide offers help generating and using an SSH key pair."03:05
man03:05
→ Squarism2 joined03:06
← yhager|away left03:07
← Squarism left03:07
← jefflyne left03:07
→ abstrusenick joined03:08
← dyng left03:08
→ jefflyne joined03:08
← jefflyne left03:08
→ jefflyne joined03:10
→ yhager|away joined03:11
RobW_ Fixed!03:13
Thanks, rubber duckies.03:13
→ bwreilly joined03:14
← guampa left03:14
→ tjj joined03:15
← bitkiller left03:16
← nwertman1 left03:17
← Zarthus left03:20
→ yacks joined03:22
→ Sakako1 joined03:23
← Sakako left03:25
RobW_ Damn, not fixed.03:25
crocket woo03:27
pdurbin, how do you use ssh-add?03:27
← edwardly left03:29
→ Sakako joined03:32
← raonyguimaraes left03:32
← Sakako1 left03:34
← dropdrive left03:35
→ minim joined03:35
→ edwardly joined03:36
→ dyng joined03:40
→ mouse_ joined03:41
← travisstaton left03:43
← MouseTheLuckyDog left03:45
→ dropdrive joined03:45
← mouse_ left03:48
→ harish__ joined03:49
← dyng left03:50
← harish_ left03:50
→ dyng joined03:50
← Earnestly left03:51
← _Harm_B_ left03:54
← rchavik left03:54
← dyng left03:55
→ _Harm_B_ joined03:55
→ milestone joined03:56
→ notjohn joined03:56
→ mizu_no_oto joined03:57
→ preflex_ joined03:58
→ Dreadnaught joined03:58
← preflex left03:58
preflex_ → preflex03:58
→ petey joined03:59
← milestone left04:00
← the_cybersphinx left04:00
→ mouse_ joined04:00
← yhager|away left04:01
→ Tommy[D]_ joined04:01
← petey left04:03
→ petey joined04:03
← kukks left04:04
← Tommy[D] left04:05
→ vkz joined04:07
← petey left04:08
→ lresende joined04:09
→ petey joined04:09
← lresende left04:10
← Naoe-Kanno left04:10
← mikepack left04:11
→ mikepack joined04:11
→ nathancahill joined04:13
← dropdrive left04:13
→ ewnd9 joined04:13
← GGMethos left04:15
← yacks left04:15
← mikepack left04:16
← PaulCapestany left04:17
← madewokherd left04:18
← raijin left04:18
→ PaulCapestany joined04:18
← amcsi left04:20
← Sakako left04:21
→ ErKa joined04:21
→ HerrK joined04:21
→ alezost joined04:22
← mizu_no_oto left04:24
→ _iron joined04:24
→ cmpera_ joined04:25
← _iron left04:25
→ _iron joined04:26
← cmpera left04:27
cmpera_ → cmpera04:27
→ pigeon joined04:27
← mj12albert left04:30
← Dougie1871 left04:31
→ dropdrive joined04:32
← Shaladdle left04:33
← _iron left04:35
→ sriracha joined04:36
→ _iron joined04:36
← _iron left04:37
→ _iron joined04:39
← dropdrive left04:40
← druonysus left04:40
→ druonysus joined04:40
← druonysus left04:40
→ druonysus joined04:40
← sriracha left04:43
→ tallest_red joined04:45
← cmpera left04:45
← crocket left04:46
← skorgon left04:47
← m0viefreak left04:48
← _iron left04:48
← fridim_ left04:48
RobW_ Fixed my use case with `--update-head-ok`, which allows fetches on the current branch04:49
← lazers left04:54
→ lazers joined04:55
→ milestone joined04:56
→ Tommy[D] joined05:00
← milestone left05:01
← minim left05:01
← Tommy[D]_ left05:02
→ CKyle joined05:04
→ sferry joined05:05
← sferry left05:05
→ sferry joined05:06
→ introom joined05:06
← vkz left05:06
← sferry left05:07
→ sferry joined05:07
← sferry left05:07
→ sferry joined05:08
← sferry left05:08
→ mizu_no_oto joined05:09
→ sferry joined05:09
→ yhager|away joined05:09
← sferry left05:09
→ sferry joined05:10
← mizu_no_oto left05:10
← sferry left05:10
→ sferry joined05:11
← sferry left05:11
→ sferry joined05:11
← sferry left05:12
→ sferry joined05:12
← sferry left05:13
→ sferry joined05:13
← sferry left05:13
→ sferry joined05:13
→ rudak joined05:13
← sferry left05:14
→ emergion joined05:14
← relipse left05:14
→ sferry joined05:14
← sferry left05:14
→ sferry joined05:15
← sferry left05:15
→ sferry joined05:15
← sferry left05:15
→ sferry joined05:15
← sferry left05:16
→ sferry joined05:16
← sferry left05:17
→ sferry joined05:17
← sferry left05:18
→ sferry joined05:18
← sferry left05:18
→ marcmarc joined05:19
→ sferry joined05:19
← sferry left05:19
→ sqlnoob joined05:20
→ sferry joined05:20
→ jruiz joined05:23
→ rurufufuss joined05:25
← marcmarc left05:26
← Squarism2 left05:29
← rudak left05:29
→ dropdrive joined05:32
← petey left05:35
→ lpapp joined05:35
→ petey joined05:36
lpapp hi, is there an option for git show HEAD not to show the content of renamed or moved files? They are noise, and it would be enough to see that they were renamed or moved, and show the inline content of the changed files instead that were not renamed, not moved?05:36
← dropdrive left05:40
← petey left05:40
← ianmcorvidae left05:40
← akurilin left05:42
→ ianmcorvidae joined05:42
← Nicholas_Rishel left05:43
← CKyle left05:51
→ skorgon joined05:54
milki lpapp: --stat05:54
o05:54
→ Jackneill joined05:54
← Jackneill left05:54
→ Jackneill joined05:54
milki theres no hybrid mode -.-05:55
→ Caritade joined05:55
→ milestone joined05:57
← jruiz left05:58
→ irqq joined05:58
lpapp milki: yeah, I would need hybrid mode.05:59
I would like to review the content, but rename and moves are just noise05:59
← tjj left05:59
← notjohn left05:59
→ dropdrive joined05:59
lpapp sounds like a feature request then!05:59
→ tjj joined06:00
← milestone left06:01
← mribeirodantas left06:01
milki lpapp: people would just tell you not to do moves and renames in the same commit >.>06:02
you are making noise for yourself06:03
← RobW_ left06:05
→ maleknet joined06:07
← dropdrive left06:07
← chuckharmston left06:08
← hyperair left06:09
→ sakkemo joined06:10
→ h0bbit_ joined06:10
lpapp milki: disagree.06:10
when you want to move and adjust the buildsystem, you do not wanna move first to break the software!06:11
so people saying that would not make sense imho. :)06:11
→ raik joined06:11
milki lpapp: disagreed. you can do these changes in a branch06:11
lpapp: and your merge commit will have all the changes in a single commit06:12
lpapp yay for extra work for no real gain!06:12
milki right now you are trying to work around git show because of the "no gain"06:12
lpapp it is like introducing a bug for extra work to be able to fix with further extra work06:12
milki i highly doubt anyone will entertain your "hybrid mode" patch06:12
and you can maintain your own fork of git all you want06:13
lpapp git simply should have a feature rather than advising the introduction of bugs with further work which can then be fixed with additional furter work.06:13
milki we call this !sausage06:13
gitinfo [!sausage_making] Some developers like to "hide the sausage making", transforming their commits before presenting them to the outside world. See http://sethrobertson.github.com/GitBestPractices/#sausage and !perfect06:13
lpapp I am sure no sane person would agree with you.06:13
yeah because people keep maintaining forks for their own features... that is agains sooo realistic.06:14
milki and !pretty06:14
gitinfo [!gitpretty] git pretty is a flowchart of how to get yourself out of trouble http://justinhileman.info/article/git-pretty/06:14
milki hm, maybe not06:14
that trigger changed06:14
lpapp again*06:14
milki looks06:14
lpapp git should simply have this useful feature06:14
milki anyways, i pretty good diffs with small overhead rather than patching git to make git show be nice06:15
milki goes back to work06:15
milki s/pretty/prefer06:15
lpapp you should not demand your preference.06:16
I mean, if you prefer the full rename mess, you can check that06:16
you might get 90% noise, or even higher ratio.06:16
that is unlikely to be productive.06:16
→ sumali joined06:17
← ewnd9 left06:18
← sumali left06:18
lpapp I would think this can be a --no-rename/remove option.06:18
→ sumali joined06:20
milki lpapp: btw, if you do submit a patch, you would submit it to the git mailing list: [email@hidden.address]06:22
→ dyng joined06:22
→ dropdrive joined06:23
→ dhruvasagar joined06:23
→ Chiko joined06:24
← dyng left06:26
→ dyng joined06:26
→ Bosox20051 joined06:27
← normanrichards left06:27
→ Targen joined06:27
lpapp milki: no, I would only request this feature. I do not have time for yet another FOSS project.06:28
→ geomyidae__ joined06:28
← geomyidae left06:29
→ hoschi joined06:29
← dropdrive left06:30
milki lpapp: then thats the mailing list to request as well06:30
← kcj left06:30
← dyng left06:31
→ ewnd9 joined06:32
→ minim joined06:34
← h0bbit_ left06:38
→ chuckharmston joined06:39
lpapp milki: I know06:40
→ _nova joined06:40
→ yacks joined06:45
→ dropdrive joined06:46
← raik left06:46
→ sumali_ joined06:47
← chuckharmston left06:47
← sumali left06:49
← ErKa left06:50
← HerrK left06:50
← minim left06:50
← Jonesboro left06:50
← sumali_ left06:50
← ewnd9 left06:51
→ Jonesboro joined06:51
→ ewnd9 joined06:51
← irqq left06:55
→ milestone joined06:57
← youlysse` left06:58
← milestone left07:02
← Jonesboro left07:03
→ mFacenet joined07:03
→ Jonesboro joined07:03
→ ewokchewy joined07:03
→ minim joined07:05
← Skrypter left07:07
→ arietis joined07:09
→ h0bbit_ joined07:10
← yhager|away left07:12
← PaulCapestany left07:13
← barneybook|2 left07:18
→ barneybook|2 joined07:18
← dreiss left07:19
← sferry left07:20
→ sferry joined07:20
← ecube left07:21
→ rat_thing joined07:21
→ ecube joined07:21
→ kongthap joined07:23
← minim left07:27
→ dreiss joined07:28
→ ngladitz joined07:29
→ babilen joined07:31
← nathancahill left07:31
← dropdrive left07:31
→ Kicer86 joined07:32
← echo-area left07:32
← rat_thing left07:33
← Kicer86 left07:33
→ rat_thing joined07:33
→ totem joined07:33
→ Kicer86 joined07:36
← babilen left07:36
← h0bbit_ left07:37
→ dyng joined07:39
→ Prazanna joined07:39
→ chuckharmston joined07:40
← diegoviola left07:42
← Goplat left07:43
→ nathancahill joined07:44
→ vurma joined07:47
← chuckharmston left07:48
← rat_thing left07:49
→ Shusshu joined07:50
→ Mars` joined07:51
← Nuck left07:51
→ PLejeck joined07:52
← Shusshu left07:52
→ milestone joined07:58
← kerrick left07:59
→ h0bbit_ joined07:59
← milestone left08:03
→ mnathani joined08:04
mnathani I cloned linus08:04
→ maqr joined08:04
mnathani I cloned linus's kernel tree08:04
← PLejeck left08:04
mnathani but I cant find the latest tag Linux 3.12-rc308:04
→ Nuck joined08:04
← bwreilly left08:04
← Caritade left08:05
iveqy mnathani: what does git tag says?08:06
mnathani v3.9-rc808:06
a whole list that goes upto that tag08:06
iveqy where did you clone from?08:06
mnathani my bad, 12 is < 908:07
so it was in the list08:07
I was simply expecting it at the bottom08:07
iveqy :)08:07
osse git: y u no have builtin version sort! :(08:07
iveqy osse: it08:09
osse: it's not easy to have a good default sort08:09
osse Maybe not08:10
I for one would expect it to stort alphabetically by default, as it does now. Just saying --sort=version would be nice :) Could yank the code from sort -v08:11
← arietis left08:12
→ asQuirreL joined08:12
→ Tritania joined08:13
iveqy osse: yeah that would have been nice. However, there's a lot of other features I would rather have since git tag | sort -v is pretty easy to do08:14
→ HelloShitty joined08:14
← totem left08:16
osse iveqy: sure. It's not a big deal. I'd rather have in-memory merges for example :P Btw, a lot of sorts don't have -v08:17
Is "in-memory merge" a thing or something I've imagined? I seem to recall a short discussion on the mailing list some time ago about the ability to merges A into B while having C checked out (or having none checked out at all, ie. a bare repo).08:19
← chrisss404 left08:20
iveqy osse: but how would you hande a conflict?08:20
seems like it would speedup merges a bit08:20
→ soa2ii joined08:21
← dyng left08:21
→ dyng joined08:21
osse maybe the conlict aspect was the nail in the coffin for the whole feature; I don't remember. Could dump them in /tmp/tmp.git-merge.DRGEAR¤#R34 like diff-/mergetool does now08:21
soa2ii Hi. I just have to commit a file to a svn (it's about 13GB :-/). Can I somehow just commit something with git svn instead of cloning it first?08:21
→ rat_thing joined08:22
charon soa2ii: you can use git svn init <args>; git svn fetch -r <revisions> to limit fetching, or patch exclusions to remove stuff, but no, you can't "blindly" commit08:23
→ Zarthus joined08:24
soa2ii charon: Hm, ok. Thanks :-)08:24
← dreiss left08:25
j416 soa2ii: would probably be easier to just use svn to commit that, I suppose08:25
less of a hassle08:25
← dyng left08:25
soa2ii j416: Hm, I'd have to look that up. But ok :-)08:26
→ dyng joined08:28
→ danicampa901 joined08:28
← sferry left08:29
← dyng left08:29
→ matanya__ joined08:29
→ dyng joined08:30
← matanya__ left08:30
→ doev joined08:33
← rat_thing left08:33
→ dyng_ joined08:34
← dyng left08:34
→ herdingcat joined08:36
lpapp milki: thanks for letting me know there is no hybrid mode just yet.08:37
← dyng_ left08:37
→ dyng joined08:38
→ dyng_ joined08:39
← dyng left08:39
→ arietis joined08:39
→ leeN joined08:39
Baluse → SugarDaddy08:39
→ chuckharmston joined08:40
SugarDaddy → Baluse08:41
← dyng_ left08:43
→ dyng joined08:44
← chuckharmston left08:45
← asQuirreL left08:45
→ rat_thing joined08:46
→ ewokchew_ joined08:46
→ Mode-M joined08:46
→ mouse__ joined08:47
← _nova left08:47
← dyng left08:48
→ dyng joined08:50
← ewokchewy left08:50
← mouse_ left08:50
← hoschi left08:51
← ewokchew_ left08:52
→ ciampix joined08:53
← mouse__ left08:53
→ chrisss404 joined08:54
→ King_Hual joined08:54
→ longbeach joined08:54
sadistik → Saddickstick08:58
→ milestone joined08:59
← dyng left08:59
→ dyng joined08:59
← introom left09:01
→ introom joined09:01
← milestone left09:03
← dyng left09:04
← thiago left09:07
→ mouse__ joined09:07
Saddickstick → gaybagel09:08
← ciampix left09:08
→ kamikaz31 joined09:09
← jefflyne left09:10
→ khmarbaise joined09:12
← harish__ left09:12
→ harish__ joined09:13
gaybagel → barking_dog09:13
→ nilg joined09:15
→ Fingolfin joined09:15
← King_Hual left09:19
→ dyng joined09:23
→ Guest20398 joined09:23
→ jnavila joined09:25
→ Misan joined09:25
← introom left09:25
← sayanee left09:25
→ sayanee joined09:26
→ introom joined09:26
← vigs left09:26
← h0bbit_ left09:27
← sayanee left09:30
← introom left09:30
← dyng left09:36
→ dyng joined09:37
→ Xethron joined09:38
← catsup left09:40
→ chuckharmston joined09:40
← dyng left09:41
→ dyng joined09:44
← realeyes left09:44
← terminalRepeat left09:45
← chuckharmston left09:45
→ dropdrive joined09:45
← dyng left09:47
→ dyng joined09:47
→ marhaban joined09:48
→ ozette joined09:49
→ Burer joined09:49
← Burer left09:49
→ Burer joined09:49
← _Harm_B_ left09:51
← rat_thing left09:51
→ _Harm_B_ joined09:51
← dyng left09:52
→ Balliad joined09:53
→ dyng joined09:53
← druonysus left09:55
← dhruvasagar left09:55
→ druonysus joined09:55
← D-Chymera left09:56
→ asQuirreL joined09:57
→ milestone joined09:59
← dsdeiz left09:59
→ King_Hual joined10:00
← el3ctron left10:00
→ dsdeiz joined10:00
→ Sniffer joined10:01
← Sniffer left10:02
→ kcj joined10:02
← Mars` left10:03
← asQuirreL left10:03
→ Mars` joined10:03
← milestone left10:04
→ edran joined10:04
→ milestone joined10:04
→ dfanjul joined10:05
← khmarbaise left10:05
→ chimay joined10:05
→ introom joined10:06
→ Anchakor joined10:07
← Mars` left10:08
→ GODDOG joined10:10
→ asQuirreL joined10:10
← asQuirreL left10:11
Anchakor hi, I need a repo mirror on my server to which I can push to, so I created a bare repo, but now I also need to view the HEAD working tree there. how do I do this? or is it possible to have a non-bare mirror repo which I can push to?10:11
→ threesome joined10:12
→ asQuirreL joined10:12
osse Anchakor: you can use git log, git show, etc. even in a bare repo10:13
Anchakor: if you need a working tree I suggest you use git archive to create a tarball and extract it somewhere10:14
Anchakor well I also need the history there10:14
osse or have a non-bare clone somewhere else on the server10:14
Anchakor I want it so I don't have to do nothing else then the push from my workstation10:14
osse since the non-bare clone will be local (with respect to the server) you can use a post-receive hook in the bare repo to update the non-bare one10:15
Anchakor ok so I will have my workstation main repo, and on server a bare repo and non-bare repo, which is updated by a post-recieve hook of the bare repo?10:16
← doev left10:16
→ roxlu joined10:16
roxlu hi10:16
osse Anchakor: that's my suggestion, yes. Don't know if it's the best way to do it but it's the best I can think of. In any case avoid pushing to a non-bare repo, because that'll just be a mess10:17
→ freestyl3r joined10:17
← Georg3 left10:17
← introom left10:17
Anchakor yeah. this seems to be quite a common config, strange that it is such complicated10:18
→ introom joined10:18
← Balliad left10:18
osse why do you need a non-bare repo with full history on the server? Don't you do all your work on your workstation ?10:19
← DonRichie left10:19
muep !deploy10:19
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/the-list-and-irc/deploy.html10:19
← lpapp left10:20
→ DonRichie joined10:20
→ Balliad joined10:20
Anchakor osse: I do, but that repo is for my supervisors to see what I am working on10:21
→ Sendoushi joined10:24
roxlu hey guys, what is a good way to deal with binary files and git repositories?10:25
← marhaban left10:25
← p3lim left10:25
→ harish_ joined10:25
osse Anchakor: heh, tell them to clone it themselves :P Or you can setup something like gitweb or redmine or whatever.10:26
← asQuirreL left10:26
← harish__ left10:26
→ _iron joined10:27
→ catsup joined10:28
← _iron left10:28
→ _iron joined10:29
← _iron left10:30
← vurma left10:30
→ smlgs joined10:30
→ marhaban joined10:30
→ the_cybersphinx joined10:30
→ _iron joined10:30
→ HumptyDumpty001 joined10:31
gitinfo set mode: +v10:32
→ Ozch joined10:32
HumptyDumpty001 Hello10:32
gitinfo HumptyDumpty001: 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.10:32
← _iron left10:33
→ Kakera joined10:33
→ mdiers joined10:34
← dyng left10:34
→ dyng joined10:34
→ sebrock joined10:36
← kcj left10:36
← jnavila left10:38
← maleknet left10:38
→ babilen joined10:38
→ arand joined10:38
→ maroloccio joined10:38
→ trandatnh joined10:38
← herdingcat left10:38
← Balliad left10:39
← dyng left10:39
→ h0bbit_ joined10:39
→ opmrcrab joined10:40
→ aChatir joined10:41
→ rednaks joined10:41
→ Balliad joined10:41
→ hoschi joined10:42
← dfanjul left10:42
Anchakor hmm my post-receive hook "cd worktree && git pull" doesn't work. any ideas?10:44
ofc it is a proper sh script with execute rights10:44
→ ahri joined10:44
ahri → Guest1929610:45
→ rouf joined10:46
→ CheckDavid joined10:46
→ Nixola joined10:47
← maqr left10:47
→ rouf_ joined10:47
Nixola hi, I'd need a little help with Git10:47
I've only started recently to use it, so I'm a bit unsure about lots of things10:48
→ _iron joined10:48
Nixola if the branch A is selected in my local repo and I do "git push github B", will the local branc A be pushed to the B branch on Github?10:48
← nemesit|znc left10:49
Nixola (github is correctly set so that I can push changes)10:49
← HumptyDumpty001 left10:49
→ HumptyDumpty002 joined10:49
opmrcrab thats basicly all i do for bitbucket, for what its worth, im relativly new myself10:49
Nixola opmrcrab: so it works?10:50
opmrcrab typicaly im doing "git push origin master" where orign is my bitbucket remote, and master is the branch im pushing to10:50
arand Nixola: Normally no, it would push local branch B to remote B, you can use A:B to push some ref to another.10:50
← rouf left10:51
→ cmpera joined10:51
Nixola git push github A:B or what?10:51
→ mklappstuhl joined10:51
→ dhruvasagar joined10:51
← trandatnh left10:52
← Guest19296 left10:52
arand If they have the same name both locally and on the remote, or if the default push branch for the local branch is set (iirc) then you can use just the single ref syntax.10:52
Nixola: That should push local A to remote B, yes.10:52
Nixola ok, thanks10:53
...damn10:54
I've been developing in branch B instead of A10:54
well, thanks anyway10:54
that'll come in handy later, probably10:54
← Baluse left10:54
Nixola bye!10:54
← Nixola left10:54
← Balliad left10:55
→ eletuchy joined10:55
→ chairabanta joined10:56
→ Baluse joined10:56
← edwardly left10:56
← Anchakor left10:56
→ edwardly joined10:56
← edwardly left10:57
← Fingolfin left10:58
→ edwardly joined10:58
← edwardly left10:59
→ edwardly joined11:00
← Baluse left11:00
→ Baluse joined11:00
← edwardly left11:01
→ rat_thing joined11:02
→ edwardly joined11:02
← introom left11:02
→ RaceCondition joined11:02
→ kcj joined11:02
← edwardly left11:03
→ edwardly joined11:03
→ penguin42 joined11:06
← rat_thing left11:06
← _Harm_B_ left11:06
→ _Harm_B_ joined11:07
osse !hooks11:08
gitinfo [!hook_pitfalls] Guidelines for writing hooks: 1. Consume all input (cat >/dev/null if you don't want it). 2. If you use any 'cd', also 'unset GIT_DIR'. 3. Don't git-pull in a hook (or any other script).11:08
→ milStaja joined11:08
→ jefflyne joined11:09
← rjc left11:10
→ nemesit|znc joined11:10
→ Balliad joined11:12
→ rjc joined11:13
← mklappstuhl left11:13
→ TomyLobo joined11:13
← Balliad left11:15
← h0bbit_ left11:17
← chimay left11:19
← rednaks left11:20
← Misan left11:21
← chairabanta left11:21
→ dyng joined11:22
→ mklappstuhl joined11:23
← Mode-M left11:24
← Sendoushi left11:24
→ vurma joined11:26
→ FernandoBasso joined11:27
← mklappstuhl left11:27
→ mklappstuhl joined11:28
← mdiers left11:28
→ harobed_ joined11:29
→ Mode-M joined11:29
← dyng left11:29
→ dyng joined11:30
← devoldmx left11:31
→ devoldmx joined11:31
→ Balliad joined11:32
← mklappstuhl left11:32
→ zhost joined11:32
← eletuchy left11:33
← pencilcheck left11:33
→ eletuchy joined11:34
← dyng left11:34
→ introom joined11:36
← Balliad left11:36
→ h0bbit_ joined11:37
← kcj left11:38
→ Balliad joined11:38
→ trandatnh joined11:40
← cads left11:40
→ dfanjul joined11:42
→ chairabanta joined11:44
← Balliad left11:45
→ Balliad joined11:47
→ dyng joined11:49
→ greengriminal joined11:50
→ mklappstuhl joined11:50
greengriminal Most likely pretty basic but I have a git commit that i'd like to remove. I have the commit id. But what is the command to remove this particular commit from the master branch11:51
Would it be `git revert <commit-id>`11:51
cbreak greengriminal: no.11:52
greengriminal: git revert would remove the EFFECT of that commit11:52
→ Cromulent joined11:52
cbreak greengriminal: that might be the best you can do (if you pushed history and don't work alone for example)11:53
or you could try to rewrite history with git rebase -i, but that is more involved11:53
← nemesit|znc left11:53
greengriminal I have a commit I'd like to revert to11:53
→ pbujok joined11:53
cbreak reverting is not to a commit11:54
bremner !revert11:55
gitinfo 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]?11:55
cbreak resetting is to a commit (but that is history rewriting)11:55
← Kwpolska left11:55
→ kcj joined11:56
← dhruvasagar left11:56
→ eletuchy_ joined11:56
← chairabanta left11:56
greengriminal I think i need to re-write history. In other words my master branch is on "d32ce7b" but I want to get rid of this and go back to "cdde845"11:57
cbreak greengriminal: then, make sure you have no uncommitted changes11:58
greengriminal: then git tag backup11:59
then git reset --hard cdde84511:59
→ blathijs joined11:59
← eletuchy left11:59
← introom left12:00
greengriminal cbreak: And "git reset --hard cdde845" will put back to that commit12:00
→ Misan joined12:00
→ introom joined12:00
← eletuchy_ left12:01
→ Kwpolska joined12:02
← aChatir left12:03
← dsdeiz left12:04
← introom left12:05
→ dsdeiz joined12:08
→ eletuchy joined12:09
← HumptyDumpty002 left12:09
← iron_houzi left12:10
→ Risto- joined12:11
→ Earnestly joined12:11
→ nardev joined12:12
← mklappstuhl left12:12
← jefflyne left12:13
← dyng left12:13
← Balliad left12:13
→ skeuomorf joined12:14
skeuomorf guys, I have a repo on github, with 3 branches, the master, one that I created, one that my coworker created, He deleted some files from the branch that I created, I did a git pull but the files aren't removed locally, how do I completely sync the remote and local branch?12:15
→ introom joined12:15
→ dyng joined12:16
→ iron_houzi joined12:16
cbreak skeuomorf: be sure you merged the right things12:16
skeuomorf: look at the commit your coworker made that removes the files12:16
did you merge that?12:16
skeuomorf I thought git pull is short for (git fetch && git merge)12:16
→ raonyguimaraes joined12:16
cbreak it is.12:17
← CheckDavid left12:17
skeuomorf ok, so I did a git pull12:17
cbreak but it pulls upstream. Did you configure upstream properly?12:17
skeuomorf file not removed from local branch12:17
cbreak so answer my questions12:17
skeuomorf I configured it using --track origin/branch12:18
cbreak did you merge that commit?12:18
(is it visible in git log branch?)12:18
→ nemesit|znc joined12:18
skeuomorf it is visible in git log12:19
iveqy skeuomorf: does your coworker has push access to your repo on github? That is are you working with the same repo?12:19
cbreak skeuomorf: then you merged it12:19
skeuomorf iveqy, affirmative12:19
cbreak skeuomorf: then it was removed12:19
→ raijin joined12:19
skeuomorf cbreak, but the file is still there12:19
cbreak skeuomorf: maybe you resolved a merge conflict in favor of your side12:19
a change/remove conflict12:19
skeuomorf: do you see the merge commit in the history?12:20
(the one from your pull?12:20
skeuomorf no12:20
no merges in git log12:20
cbreak then it was fast forward?12:20
git log --stat12:20
you see his commit12:21
after that, the files must be recreated12:21
because you said that commit removed them12:21
skeuomorf yeah12:21
they are recreated12:21
cbreak then delete them again if you want them gone.12:22
← grilix left12:22
skeuomorf cbreak, awesome, thanks a dozen :)12:22
→ grilix joined12:22
← Cromulent left12:23
→ _nova joined12:23
→ CheckDavid joined12:24
← sqlnoob left12:24
← xobb left12:25
→ subhojit777 joined12:25
← dyng left12:26
→ dyng joined12:27
← JPeterson left12:30
← harobed_ left12:31
← dyng left12:31
→ mklappstuhl joined12:31
← raonyguimaraes left12:32
← Kakera left12:35
greengriminal why is it if I do "git rebase -i d32ce7b^" delete the offending commit and then :wq and it says "Nothing to do"12:37
Then when I view go back into that same commit using the git rebase -i d32ce7b^ the line is still there. Its as though it doesn't save my changes12:38
cbreak greengriminal: maybe the list is empty12:38
greengriminal: rebase -i requires a non-empty work list12:38
← kcj left12:39
greengriminal cbreak: What does that mean?12:39
cbreak greengriminal: rebase -i. do you know what it does?12:40
← mklappstuhl left12:40
→ Guest68708 joined12:40
→ irqq joined12:40
greengriminal cbreak: it allows you to make change from one branch into another12:40
cbreak it cherry-picks a series of commits onto the chosen base, optionally stopping in between for editing or changing commit messages12:41
it does that by using a list of work items, which it presents to the user (you)12:41
if that list is empty, it will not do anything12:41
→ chuckharmston joined12:44
← thierryp_ left12:44
→ thierryp joined12:45
→ tmcmahon joined12:46
← pbujok left12:46
← amunoz left12:47
skeuomorf I added a file to my master branch, Now I want to add that file to another branch, I checked out the 2nd branch and did git merge master, resolved some conflicts regarding files that has already been there, but the file wasn't merged12:47
→ Amandil joined12:48
← greengriminal left12:48
iveqy skeuomorf: have you solved the conflict and commited?12:48
skeuomorf yes12:48
← chuckharmston left12:48
cbreak skeuomorf: after a merge, all commits you merged are part of history12:48
→ hyperair joined12:48
cbreak if one of those commits added a file, then you either have that file, or a later commit removed it again.12:48
skeuomorf hmmm12:49
I will look at the git log again12:49
cbreak you can do git log --stat -- file if you want12:49
← thierryp left12:50
cbreak (beware of merge commits)12:50
skeuomorf hmmm12:50
→ lotus joined12:51
lotus hey, I'm working on a programming project and I want to upgrade the old core CMS that the project was built around. However, the CMS files have been hacked a bit. The versions aren't too far from eachother and I'm wondering if I can just use a diff of the current directory against a clean version of the old CMS. Then use that diff, somehow, to apply the same changes to the new version of the core CMS files12:51
skeuomorf cbreak, Can I send screenshots on #git?12:51
lotus I've made a branch off dev and I'm thinking if I just unzip the clean version of the old CMS ontop of the current hacked version, then make a reverse diff?12:51
cbreak skeuomorf: heard of pastebin?12:51
→ mklappstuhl joined12:52
lotus I'm just wondering what the best way to do this would be using the git diff tool?12:52
skeuomorf cbreak, yeah, yeah, I meant printscreen, image12:52
cbreak what's the point of that?12:52
there are image upload pages too though12:52
skeuomorf show something from gitg12:52
by can I send it, I meant it's not against channel rules, correct?12:53
cbreak no12:53
links are OK.12:53
← tmcmahon left12:53
skeuomorf ok12:54
→ Kakera joined12:54
← twoface88 left12:54
→ joelmo joined12:55
skeuomorf cbreak, here are the commits http://i.imgur.com/XQUDhPJ.jpg12:55
cbreak, the file local.py has been added to master, I want to add it to the "social" branch12:55
cbreak everything on master is there already12:56
you're done.12:56
joelmo i have tracked https repositoreis that requires password, is there any way i could have a password manager for these so that i only need to type in that password say twice a day?12:56
→ m0ltar joined12:56
skeuomorf cbreak, the file local.py is not availabe on social12:56
cbreak skeuomorf: then you deleted it again12:56
skeuomorf: add it again if you want to have it.12:56
skeuomorf cbreak, No, didn't delete it, but how do I add it?12:56
cbreak skeuomorf: in the merge commit, you likely deleted it as result of the resolution12:56
skeuomorf: of course you deleted it.12:56
skeuomorf cbreak, hmm, ok12:56
cbreak skeuomorf: look at the diff between the merge and master12:57
skeuomorf: git diff master..social12:57
skeuomorf: you can do git checkout master -- local.py12:57
that'll get you master's local.py file12:57
then commit it12:57
skeuomorf git checkout master -- local.py showed me "error: pathspec 'local.py' did not match any file(s) known to git."12:58
cbreak then master didn't have that file.12:59
skeuomorf cbreak, I checked out master, navigated to the folder, it's there12:59
cbreak, hold on13:00
cbreak no folder...13:00
you told git it's in the root...13:00
skeuomorf cbreak, I edited git checkout master -- directory/local.py, that's when I got the error13:00
but I think I know my problem13:00
→ dyng joined13:00
skeuomorf the local.py file is in .gitignore13:00
← threesome left13:01
skeuomorf stupid!13:01
cbreak that's irrelevant13:01
.gitignore doesn't affect tracked files13:01
← edwardly left13:01
→ asQuirreL joined13:01
skeuomorf I am confused right now13:01
→ edwardly joined13:01
← Prazanna left13:02
← dyng left13:02
← m0ltar left13:02
← edwardly left13:02
→ dyng joined13:02
→ edwardly joined13:02
← nfroidure left13:03
← edwardly left13:03
→ edwardly joined13:03
→ thierryp joined13:04
← edwardly left13:04
→ edwardly joined13:04
→ GT4066 joined13:06
→ tobscore joined13:06
← edwardly left13:06
← thierryp left13:06
← dyng left13:07
skeuomorf cbreak, I fixed it, thanks a lot for your help, sorry about my noobiness :)13:07
→ thierryp joined13:07
→ edwardly joined13:07
cbreak np13:07
→ zheng joined13:07
→ mixel joined13:07
← asQuirreL left13:07
→ doev joined13:07
← edwardly left13:08
BinGOs .version13:08
gitinfo BinGOs: .version: still at 1.8.4, not updating topic.13:08
BinGOs lies13:08
→ edwardly joined13:09
← edwardly left13:10
→ dyng joined13:10
→ edwardly joined13:10
→ SHyx0rmZ joined13:11
← thierryp left13:11
← edwardly left13:12
→ stamina joined13:12
→ edwardly joined13:12
← edwardly left13:14
→ asQuirreL joined13:14
← dyng left13:14
→ edwardly joined13:14
→ dyng joined13:14
← edwardly left13:15
→ Scorchin joined13:15
→ edwardly joined13:15
← arietis left13:15
← edwardly left13:16
→ edwardly joined13:16
→ koda joined13:17
Scorchin What are the best resources to read to learn more about how git and gpg signing works?13:18
← dyng left13:18
Scorchin Even better if they explain the shortcomings of using git and gpg13:18
→ dyng joined13:18
← _Harm_B_ left13:20
cbreak shortcommings?13:20
→ _Harm_B_ joined13:21
← koda left13:22
← dyng left13:23
→ dyng joined13:23
→ koda joined13:23
← introom left13:23
← dyng left13:23
→ dyng joined13:24
← Davey left13:25
← lotus left13:27
← trandatnh left13:28
→ bsaboia_ joined13:28
← Guest68708 left13:29
← asQuirreL left13:29
← Risto- left13:30
→ introom joined13:30
MrDHat → MrDHat|offline13:30
← nemesit|znc left13:31
← dfanjul left13:32
→ dfanjul joined13:32
→ tagava joined13:34
← Edgan___ left13:34
→ Guest68708 joined13:34
← stamina left13:35
→ Balliad joined13:36
→ Balliad_ joined13:37
→ thierryp joined13:38
→ szx joined13:39
← dyng left13:40
→ dyng joined13:40
→ m0viefreak joined13:41
← Balliad left13:41
→ travisstaton joined13:41
→ trandatnh joined13:42
→ whowantstolivef3 joined13:43
← cmpera left13:43
← Guest68708 left13:44
→ chuckharmston joined13:44
f8ld → f8l13:44
Scorchin cbreak: yeah, I'm thinking about using gpg to sign commits to know that they were created by they claim to be13:45
← dyng left13:45
cbreak git doesn't add shortcommings.13:45
Scorchin cbreak: one of the failings I have to deal with, is that transferring files between sites can only be done through HTTP (no SSL) and I want to guarantee someone can't change the repo in transfer13:45
cbreak it's just a signature of a tag (or commit)13:45
Scorchin okay13:46
← thierryp left13:46
cbreak you can look up the involved parts.13:46
← whowantstolivefo left13:46
cbreak potential weaknesses are sha1 (which is used for all git hashes), and the usual GPG weaknesses such as bad keys, insecure storage of private keys, ...13:47
← mdupont left13:47
← zheng left13:47
← King_Hual left13:47
cbreak sha1 is kind of nearing the end of its lifetime13:47
with gpg, if you use a 2048bit or bigger RSA key, you might be OK for the next few years13:47
Scorchin great, thanks13:47
as long as I can use signed tags/commits to guarantee the integrity of my repos when they're in transfer :-)13:48
← szx left13:50
← asakura left13:51
← chuckharmston left13:51
← trandatnh left13:53
→ dyng joined13:53
← Mode-M left13:54
→ nemesit|znc joined13:55
← salamanderrake left13:56
← dfanjul left13:56
→ ferdev joined13:57
← ferdev left13:57
→ King_Hual joined13:58
→ dfanjul joined13:59
→ ehmatthes joined14:01
cbreak guarantee within limits.14:02
← emergion left14:04
← koda left14:06
← rurufufuss left14:06
→ trandatnh joined14:07
→ ciampix joined14:08
→ bodom joined14:10
← milStaja left14:10
← mklappstuhl left14:10
← edran left14:11
← dyng left14:12
→ jefflyne joined14:12
← dfanjul left14:12
→ dyng joined14:12
→ amcsi joined14:13
← dyng left14:13
← Balliad_ left14:13
→ dyng joined14:13
← nemesit|znc left14:13
→ koda joined14:14
barking_dog → barking_cat14:15
→ Balliad joined14:15
→ fridim_ joined14:15
← koda left14:17
→ koda joined14:17
→ mklappstuhl joined14:18
← RaceCondition left14:18
← dyng left14:18
← Milossh left14:18
→ dyng joined14:19
→ RaceCondition joined14:19
→ rurufufuss joined14:19
← introom left14:20
→ nemesit|znc joined14:21
→ dyng_ joined14:21
← dyng left14:21
→ staafl joined14:23
→ Milossh joined14:24
← dyng_ left14:25
→ dyng joined14:25
→ jaskirat joined14:27
→ sayanee joined14:27
gitinfo set mode: +v14:28
gitinfo set mode: +v14:29
← NicolasTarzia left14:29
→ petey joined14:29
← dyng left14:29
→ chairabanta joined14:30
← cagedwisdom left14:30
← Fandekasp left14:30
→ NicolasTarzia joined14:30
→ Destos joined14:30
→ lite_ joined14:30
jaskirat y14:31
← abstrusenick left14:31
canton7 z14:31
← jaskirat left14:33
→ pantsman joined14:34
← pantsman left14:34
→ pantsman joined14:34
→ dyng joined14:34
→ emergion joined14:34
← dyng left14:35
→ dyng joined14:36
bremner !why14:36
gitinfo Why? Because screw you, that's why.14:36
→ whowantstolivef2 joined14:37
← nemesit|znc left14:39
→ jaskirat joined14:39
← jaskirat left14:39
← whowantstolivef3 left14:39
→ nemesit|znc joined14:40
→ bitkiller joined14:40
← dyng left14:41
← RaceCondition left14:43
→ kukks joined14:44
← hoschi left14:44
→ dfanjul joined14:44
→ RaceCondition joined14:44
← sayanee left14:45
→ sayanee joined14:45
→ tmcmahon joined14:47
→ arietis joined14:47
→ salamanderrake joined14:47
→ LuxuryMode joined14:49
← tmcmahon left14:49
→ Dougie187 joined14:50
→ cmpera joined14:50
← sayanee left14:50
← marhaban left14:50
→ yhager|away joined14:52
→ d3c joined14:53
← trandatnh left14:55
← bodom left14:56
→ notjohn joined14:56
← mklappstuhl left14:57
→ dhruvasagar joined14:59
→ taptapdan joined15:00
→ tho joined15:02
← dfanjul left15:02
→ CKyle joined15:03
→ JPeterson joined15:04
→ mklappstuhl joined15:04
→ xxpor_ joined15:05
← nathancahill left15:05
← koda left15:06
← rurufufuss left15:06
← xxpor left15:08
→ PeterBooker joined15:10
→ dfanjul joined15:10
← h0bbit_ left15:12
→ nathancahill joined15:12
← nathancahill left15:14
← CKyle left15:15
← mklappstuhl left15:15
← jefflyne left15:16
← arietis left15:17
→ mklappstuhl joined15:17
← LuxuryMode left15:17
→ sriracha joined15:17
→ aslaen joined15:18
→ thierryp joined15:18
→ wrouesnel1 joined15:18
→ chimay joined15:18
→ rurufufuss joined15:19
← yhager|away left15:20
← thierryp left15:20
→ arietis joined15:21
→ vkz joined15:21
→ twoface88 joined15:22
→ thierryp joined15:23
← dsdeiz left15:23
← freestyl3r left15:23
→ LuxuryMode joined15:24
← twoface88 left15:24
Ivo I have 3 commits going out to a pr, and i want to back out/store/stash the outer two and leave the middle two, what's the path of least resistence for that15:25
→ stamina joined15:26
→ byprdct joined15:27
← byprdct left15:27
→ v1n joined15:27
→ freestyl3r joined15:28
iveqy Ivo: why do you wan't to do that?15:28
Ivo: do you wanna remove those two from the PR?15:28
← FernandoBasso left15:29
milki Ivo: !fixup15:30
gitinfo Ivo: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!15:30
→ h0bbit_ joined15:31
→ Peter joined15:31
← bsaboia_ left15:31
← dfanjul left15:31
← JPeterson left15:31
Peter → Guest5081015:31
→ pdx joined15:35
→ geak joined15:35
→ CKyle joined15:37
gitinfo set mode: +v15:37
Guest50810 asdf15:38
diff --git a/thirdfile.txt b/thirdfile.txt15:38
index 7caac66..f6eb45c 10064415:38
Binary files a/thirdfile.txt and b/thirdfile.txt differ15:38
This is all I get with diff. How do I get to show the lines that have been changed?15:38
→ dfanjul joined15:38
iveqy diff --git? What diff program is that?15:39
Ivo iveqy: yeah basically. I ended up going back, branching, and cherrypicking commits15:40
Guest50810 just git diff on the command line in windows 715:40
← Ozch left15:40
iveqy Guest50810: no... you runned the diff program15:41
← gitpicker left15:42
Guest50810 What does that mean?15:42
← lite_ left15:42
→ gitpicker joined15:43
iveqy Guest50810: diff is one program, git is an other program15:43
try git diff instead15:43
← Burer left15:43
Guest50810 That what's I did15:44
iveqy Guest50810: you said you did: diff --git15:44
what did you do?15:44
← CKyle left15:44
Guest50810 I did git diff and got this back:15:46
diff --git a/thirdfile.txt b/thirdfile.txt15:46
index 7caac66..f6eb45c 10064415:46
Binary files a/thirdfile.txt and b/thirdfile.txt differ15:46
→ nathancahill joined15:46
← nathancahill left15:46
→ nathancahill joined15:48
← nathancahill left15:48
iveqy Guest50810: okay, so git thinks that your files are binary, do you have any .gitattributes?15:48
→ threesome joined15:49
Guest50810 I don't know. I'm trying to figure out how to tell15:49
→ hrefchef joined15:51
← dfanjul left15:51
← nemesit|znc left15:52
→ dfanjul joined15:52
Guest50810 How do I see/edit attributes?15:53
cbreak attributes?15:53
← mixel left15:53
← chimay left15:53
cbreak Guest50810: you mean man gitattributes?15:53
gitinfo Guest50810: the gitattributes manpage is available at http://jk.gs/gitattributes.html15:53
cbreak Guest50810: with a text editor.15:53
← pantsman left15:54
→ mixel joined15:55
Guest50810 I don't have any attributes file15:55
→ chuckharmston joined15:56
grawity then you don't have any attributes yet, either15:57
← threesome left15:58
→ twoface88 joined15:58
cmn are you sure the files *are* text?15:58
→ mfulz joined15:58
→ Squarism joined15:59
cmn does 'file' agree it's a plain text file?15:59
← Scorchin left15:59
iveqy Guest50810: what does: file thirdfile.txt show you?15:59
← CheckDavid left16:00
→ joshsmith joined16:01
j416 iveqy: that's normal output from git diff16:01
→ a-l-e joined16:02
j416 (diff --git a/path b/path)16:02
← emergion left16:02
→ JPeterson joined16:02
iveqy j416: yes I know, I missunderstood him that that was the command he runned16:02
j416 Guest50810: I bet your file is in UTF-16. Use UTF-8.16:03
→ ThomasCompix joined16:04
→ Edgan___ joined16:04
Guest50810 file: command not found16:05
iveqy Guest50810: how about: cat thirdfile.txt16:05
→ Scorchin joined16:05
→ shuaib joined16:06
j416 sigh. UTF-16. *hint*16:06
:)16:06
Scorchin Is it possible for someone to spoof signing of a git tag? How does the signing of commits work in git? How hard is it for others to tamper with?16:06
→ shoshin joined16:06
j416 Microsoft loves UTF-16.16:06
→ victorbjelkholm joined16:06
cmn Scorchin: it works exactly the same as every other time you sign anything16:07
if someone else has the private key, they can sign too16:07
there's nothing git specific aboutit16:07
Guest50810 j416 I heard you, I takes me a long time to figure anything out.16:07
→ shihan joined16:08
iveqy Scorchin: signing a commit is not the same as signed-off: ...16:08
← mklappstuhl left16:08
Scorchin iveqy: I got that bit. I'm more trying to work out how someone might break the use of signed commits.16:09
→ mklappstuhl joined16:09
grawity Git uses PGP for signed tags and commits – so it depends on the security of PGP itself16:09
cbreak Scorchin: easiest method: remove the signature.16:10
grawity (and on the security of the SHA1 hash)16:10
→ CKyle joined16:11
→ nathancahill joined16:11
→ oriba joined16:12
← nathancahill left16:12
j416 Guest50810: wild guess, you're using visual studio or similar16:13
Guest50810 I was able to get the diff details by changing the file to utf-816:13
j416 :)16:13
Guest50810 I'm using Ultraedit16:13
j416 ah16:13
cmn right, you make the file text, you can get a diff16:13
→ CheckDavid joined16:13
j416 good to know that ultraedit uses utf-16 by default (?) too16:13
→ hoschi joined16:14
j416 cmn: hm well, utf-16 is also text :)16:14
Guest50810 But if my files are no utf-16 doesn't that mean I can't use some unicode charactars?16:14
j416 git just doesn't get that16:14
→ rudak joined16:14
cmn j416: it's not16:14
→ buhman joined16:14
← dhruvasagar left16:14
buhman how do I create a new branch that has no parent commit?16:14
j416 Guest50810: you can use all unicode characters16:14
Guest50810: UTF-8 is just a different, more common, representation of Unicode16:15
→ nemesit|znc joined16:15
neal__- → neal__16:15
Guest50810 Ok thanks.16:15
cbreak UTF-8 can represent more characters than UTF-16, in theory16:15
cmn huh?16:15
cbreak in practice, those characters were removed from the Unicode standard16:15
cmn are you thinking of UCS-2?16:15
→ tatsuya_o joined16:15
cbreak cmn: no16:15
j416 what characters were removed?16:16
j416 lost16:16
← arietis left16:16
cmn what would utf-8 be able to represent that utf-16 can't?16:16
cbreak originally, utf-8 was planed to be up to 6 bytes per glyph16:16
it is now only up to 516:16
416:16
→ jruiz joined16:16
iveqy Guest50810: my guess is that msys doesn't really like utf-16...16:17
cbreak cmn: the UTF-16 encoding is weird. It uses special characters for multi-16-bit groups16:17
j416 iveqy: it doesn't matter. Git doesn't know how to diff UTF-16, even on not-windows16:17
← hyperair left16:17
cbreak I forgot the details but it was quite different from utf816:17
iveqy j416: really? I thought so... :(16:18
→ Saurabh joined16:18
cmn but you can still have more than 16 bits per codepoint16:18
cbreak yes.16:18
← aslaen left16:18
cbreak but it's only enough for all of the existing unicode code points16:18
not for all of the unicode code points that existed before they removed them16:18
j416 iveqy: I haven't experimented much, maybe there is some setting. You could use your own diff driver to convert it first I suppose.16:18
cbreak as I said above: not a practical problem16:18
j416 cbreak: interesting16:19
cmn that sounds for the par for unicode16:19
j416: there is no setting; UTF-16 embeds zeroed bytes16:19
← mixel left16:20
→ CycoCid joined16:20
← nicxvan_ left16:21
→ raonyguimaraes joined16:21
← Dreadnaught left16:22
← ehmatthes left16:22
→ mizu_no_oto joined16:23
← Vampire0 left16:23
j416 cmn: what's a zeroed byte?16:23
cmn a byte that is zero16:23
← chairabanta left16:23
← h0bbit_ left16:23
cmn which hits git's heuristic for binary data16:24
j416 aah16:24
clever16:24
could easily check for a BOM, though16:25
UTF-16 without BOM is wrong anyway16:25
might as well treat it as binary16:25
→ arietis joined16:26
← dfanjul left16:26
← shinnya left16:26
cmn but that would involve adding a shitton of extra handling inside git16:26
→ frgtn joined16:27
j416 just read the first two bytes before diffing?16:27
j416 shrugs16:27
j416 I don't know much about the internals16:27
cbreak just add a new diff driver for .utf1616:27
easy.16:27
cmn the internals are made for utf-816:27
j416 that's what I thought16:28
@ driver16:28
cmn reading the BOM is the least16:28
cbreak you'll have to rename your text files, but that shouldn't be an issue16:28
j416 rename.. nah16:28
well16:28
← mfulz left16:28
j416 if the project uses both UTF-8 and UTF-16, then it's.. fail lol16:29
just .gitattribute whatever file16:29
← Saurabh left16:30
cbreak j416: renaming is legit16:31
j416: windows uses file extensions to designate file content16:31
OS X too, nowadays.16:31
no one knows what linux uses... :)16:31
(probably nothing at all)16:31
← nemesit|znc left16:31
j416 magic :)16:31
grawity a mix of extensions and magic16:31
cbreak git apparently fails at magic16:31
so you'll have to rely on extensions :)16:32
→ nemesit|znc joined16:32
← ozette left16:32
→ dfanjul joined16:32
hrefchef The linux command "file" determines filetype. Extensions usually determine icons in graphical file browsers.16:34
→ Vampire0 joined16:34
→ emergion joined16:34
skorgon and 'file' implementes magic16:34
ojacobson The lookup tables and rule lists file(1) uses are (A) very long and (B) prone to fucking up in hilarious ways16:35
← mklappstuhl left16:36
→ hyperair joined16:36
← v1n left16:37
← skeuomorf left16:37
→ VictorCL joined16:37
→ mklappstuhl joined16:37
→ mribeirodantas joined16:37
→ diget joined16:38
← jday_ left16:40
Guest50810 When using save as for local files in Ultraedit it uses whatever encoding was selected the last time save as was used. I probably selected utf-16 once many years ago but I rarely save files locally so I never noticed it since them16:40
.16:40
← notjohn left16:41
← arietis left16:42
ojacobson whack the file with iconv and move on :)16:42
← wrouesnel1 left16:43
← rue_XIV left16:43
→ rue joined16:44
→ xobb joined16:44
Guest50810 UTF-16 is a historical accident that persists mainly due to inertia. UTF-16 has no practical advantages over UTF-8, and it is worse in some ways.16:44
http://benlynn.blogspot.com/2011/02/utf-8-good-utf-16-bad_07.html16:44
I never knew that.16:44
← mklappstuhl left16:44
← vkz left16:44
→ mklappstuhl joined16:45
→ rue_XIV joined16:45
ojacobson Written by someone whose entire universe is European and North American languages first, everything else second.16:46
← dfanjul left16:46
← stfn left16:46
← pdx left16:46
→ dfanjul joined16:46
→ stfn joined16:46
→ Ozch joined16:47
← rue left16:47
← _Harm_B_ left16:50
→ _Harm_B_ joined16:51
← Baluse left16:53
← nilg left16:54
cbreak Guest50810: that UTF-16-is-countable person in the comments seems clueless16:54
→ vkz joined16:55
→ DaSourcerer joined16:55
cbreak Guest50810: apparently he hasn't even heard of combining characters16:55
← jfinn left16:55
cbreak unicode is a complex thing.16:56
ojacobson No, it's worse16:56
→ Etheco joined16:56
→ minim joined16:56
ojacobson his first post talks about "if you ensure there are no Surrogate Chars" which is, uh, the point of UTF-16, so16:57
if you ensure that your string is valid UCS-2, then yes, you can use UCS-2 techniques on it16:57
DaSourcerer I've got a question regarding git subtrees: Is it possible to only import a given subdirectory of a remote repository into my project?16:57
ojacobson well done son16:57
→ chairabanta joined16:57
cbreak DaSourcerer: not without rewriting history.16:57
git history is atomic and can not be split (without creating new history)16:58
or breaking sha1 :D16:58
DaSourcerer Hm, okay16:58
→ skoop joined16:58
DaSourcerer I just thought there were a way to pollute my local repo with less clutter. Oh well.16:59
Anyway, thanks for you help17:00
← PParth left17:01
← melee left17:01
→ chisler joined17:02
→ jnavila joined17:02
→ pbujok joined17:03
→ jday_ joined17:03
← unreal left17:04
→ matanya joined17:05
→ mixel joined17:05
→ unreal joined17:05
← taptapdan left17:06
← dfanjul left17:06
→ dhruvasagar joined17:06
→ jwmann joined17:08
← Etheco left17:08
→ Etheco joined17:08
← kamikaz31 left17:09
← Soliton left17:09
→ Soliton joined17:11
→ jefflyne joined17:15
→ mikepack joined17:15
← dhruvasagar left17:16
← babilen left17:17
← tatsuya_o left17:18
→ tatsuya_o joined17:19
← shoshin left17:21
← jnavila left17:21
← doev left17:22
→ thiago joined17:22
→ v0n joined17:22
← mixel left17:23
← tatsuya_o left17:23
→ jruiz_ joined17:23
← ewnd9 left17:24
jwmann So I've attempted to rename the directory that my submodule resides in (doing a 'mv' command and relinking it in .gitmodules) but it didn't quite work out. Is there somewhere that I've missed in relinking it?17:25
← petey left17:25
← maroloccio left17:25
→ petey joined17:25
← jruiz left17:25
→ rat_thing joined17:25
← petey left17:26
→ petey joined17:26
→ LindsayMac joined17:26
LindsayMac hey guys. Ive been using github for some time on my machine with no problems. Today I open terminal to try and install node.js frm a repo and i'm told that git is not a command (i am on a mac)17:27
what in the world am i doing wrong?17:27
→ RobW_ joined17:27
SamB did you recently upgrade anything?17:28
for example, xcode?17:28
← subhojit777 left17:28
LindsayMac I am actually installing XCode now17:29
per the node.js install instriuctions17:29
← nemesit|znc left17:29
→ daanavitch joined17:29
→ shoshin joined17:30
← matanya left17:31
LindsayMac woah. why is xcode 2GB?!17:31
SamB so did you perhaps have a half-installed copy of git from XCode?17:31
ojacobson It's the entire dev kit and SDK for OS X.17:31
And iOS.17:31
LindsayMac AHH ok17:31
so yea, uits half installed17:32
← mklappstuhl left17:32
LindsayMac currently. I didnt know it was SO DAMN BIG17:32
ojacobson You can, of course, install git on its own. The version in XCode is a bit out of date, though not badly.17:32
→ nemesit|znc joined17:32
LindsayMac I had it installed on it's own17:32
ojacobson XCode also gets you important things like clang and gcc so I'd say let it rice17:32
→ dfanjul joined17:32
ojacobson ride17:32
← rouf_ left17:32
LindsayMac but found a bug http://www.hongkiat.com/blog/mountain-lion-git-fix/17:32
SamB LindsayMac: what happens when you run "which git"?17:32
(also, where had you installed git?)17:32
← barking_cat left17:32
← freestyl3r left17:33
LindsayMac i get 'git is not a command17:33
→ freestyl3r joined17:33
LindsayMac where.. hmm that is a good damn question. It was a long time ago17:34
i'm pretty sure inthe user root17:34
→ barking_cat joined17:34
→ arietis joined17:35
cbreak LindsayMac: get a mac17:35
LindsayMac i am on a mac17:36
cbreak then get Xcode17:36
then install the command line tools17:36
-> instant git17:36
LindsayMac did you not read anything that was said?17:36
→ axs joined17:36
LindsayMac just curious, because if you had, you would see i am in the process of installing xcode17:36
cbreak then git will be a valid command.17:37
LindsayMac I' not about to explain this again for oyu17:37
sorry17:37
← |Freeman| left17:37
LindsayMac SamB: thank you. hopefully letting this install will finish the job17:37
← mFacenet left17:38
→ iderik joined17:38
→ mklappstuhl joined17:38
SamB cbreak: the current theory is that the in-progress xcode install has somehow disrupted the previous git install ...17:40
← shoshin left17:40
cbreak Xcode installs to /usr/bin, normal git is likely /usr/local/bin17:41
← v0n left17:42
← petey left17:42
← frgtn left17:43
→ frgtn joined17:43
→ petey joined17:43
→ Skrypter joined17:44
← Bosox20051 left17:44
→ koda joined17:44
LindsayMac installing xcode command line tools17:44
That SHOULD do it17:44
love how this happens during a hackathon with a deadline17:46
→ madewokherd joined17:46
→ mixel joined17:46
→ devoldmx3 joined17:46
← petey left17:47
LindsayMac awesome. UP AND RUNNING AGAIN!17:47
makes me happy17:47
← Skrypter left17:48
← frgtn left17:48
→ petey joined17:48
← devoldmx left17:48
→ Skrypter joined17:49
← Skrypter left17:49
→ Skrypter joined17:50
← jruiz_ left17:51
→ rouf joined17:51
→ mFacenet joined17:52
← DaSourcerer left17:53
LindsayMac wow installing node.js was worryingly easy :)17:53
cbreak java script...17:53
← thierryp left17:53
→ thierryp joined17:54
→ thierryp_ joined17:54
← thierryp left17:54
← Jonesboro left17:55
← axs left17:55
← tho left17:55
LindsayMac yes17:55
→ zarubin joined17:55
← Etheco left17:56
→ Etheco joined17:56
← looper left17:56
→ Nicholas_Rishel joined17:57
→ devoldmx joined17:57
← devoldmx3 left17:57
← Etheco left17:57
→ rgr joined17:57
→ Etheco joined17:57
← minim left17:58
→ tatsuya_o joined17:58
← Guest20398 left17:59
→ jceb joined17:59
roxlu hi guys, what is the best way to deal with static libraries and git? Do I just add them to the repos?17:59
← mizu_no_oto left17:59
jceb → Guest711917:59
→ asakura joined18:00
LindsayMac you can add and then ignore i believe18:00
that way you aren't always messing with them18:00
rgr Should I use the pre-commit hook to run a script to update some files that get included in that git invocation? (in this case I want to run a script to pack and obfuscate a suite of js files).18:00
roxlu LindsayMac: thanks18:00
canton7 rgr, generally you shouldn't track that stuff in a git repo18:00
LindsayMac or you can use a CDN i suppose18:00
roxlu ah yeah I thought about that18:00
← dfanjul left18:01
canton7 make creating them part of your deploy process18:01
rgr canton7: I want to so the git push and resultnt side server pull ensures its all up to date and "in sync".18:01
→ shoshin joined18:02
← emergion left18:02
canton7 rgr, !deploy18:02
gitinfo rgr: 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/the-list-and-irc/deploy.html18:02
← sebrock left18:02
canton7 don't use 'git pull' for deploying18:02
roxlu canton7: yeah, thought about that.. but compiling the libraries is quite a process for people who aren't experienced with it18:02
canton7 make it so that a 'git push' runs a post-receive hook (see the link in the trigger), which creates the packfiles, and deploys the lot18:03
→ sm0ke joined18:03
roxlu canton7: can you explain that a bit?18:03
→ v0n joined18:04
canton7 oops, I'm getting rgr and roxlu confused18:04
sm0ke hey guys so i had this fork on which i was changing code .. and i made a commit to it bfore merging upstream and pushing it my local fork..18:04
← v0n left18:04
roxlu haha18:04
sm0ke now how do i handle this18:04
← Guest7119 left18:04
rgr works fine for years - but I'll lookin into why it/I shouldnt. I want the packed files local and remote you see. I also dont want to have to install the packing routines on the remove foreign owned web server. So If I need them locally as well as remotely it's fairly straightforward to just include them in the repo. They're not large. thanks anyway.18:05
← Chiko left18:05
← rgr left18:05
→ FernandoBasso joined18:05
sm0ke if i submit a pull request it will be a mess18:05
canton7 1) they're an absolute pig to merge, 2) you've got two different files holding the same content, and they're gonna get out of sync18:05
← tatsuya_o left18:06
→ tatsuya_o joined18:06
← devoldmx left18:07
→ Advocation joined18:08
→ axs joined18:08
→ devoldmx joined18:08
→ tvw joined18:08
sm0ke so can i undo a 'git merge upstream/master' and then undo my commit to local changes?18:10
← tatsuya_o left18:11
← Droolio left18:11
→ v0n joined18:12
skorgon sm0ke: i guess the question is, what do you try to achieve? merging upstream into your branch shouldn't break anything, imho, but might have been useless.18:12
sm0ke skorgon: exactly its harmless..but the git summary of pull request will look ugly..containing all the changes which have been made since i last did a merge with my current commit18:14
ideally i should have done a merge upstream before commiting and pushed to my fork18:14
then commit my changes18:14
→ dfanjul joined18:15
sm0ke i just messed up..there has to be a way out right?18:15
→ peroht joined18:15
skorgon sm0ke: IMHO, you should not have merged upstream at all. or you should have rebased your changes on top of upstream, in case there is a conflict18:15
LindsayMac has anyone found an issue with permissions denied when trying to pull a repo?18:15
She has tried this http://stackoverflow.com/questions/13195814/trying-to-git-pull-with-error-cannot-open-git-fetch-head-permission-denied18:15
skorgon sm0ke: check out the state you want from the !reflog18:16
gitinfo sm0ke: The git reflog (`git log -g`) records the SHAs of your HEADs for 2+ weeks. `git checkout -b myrestore OLDSHA` and `git reset --hard OLDSHA` will relink to that state via a new and current branch respectively, see http://sethrobertson.github.com/GitFixUm/ for full details. WARNING: reset --hard will trash any uncommitted changes! Visualize with: gitk --all --date-order `git log -g --pretty=%H`18:16
→ asQuirreL joined18:16
LindsayMac tried to do the chown thing and i get chown: username: illegal group name18:16
An_Ony_Moose LindsayMac: which OS are you on?18:17
← mklappstuhl left18:17
LindsayMac I am asking for someone, but she is on a mac18:17
cmn chown is the right thing (and stop doing stuff as root)18:17
An_Ony_Moose LindsayMac: if they have admin rights, try using username:staff instead18:18
cmn what is the *actual* command that's causing issues?18:18
An_Ony_Moose so chown -R username:staff .git/18:18
LindsayMac An_Ony_Moose: she said she is logged in as admin on your computer.18:18
adaptr cat /etc/issue18:18
An_Ony_Moose LindsayMac: chown -R username:staff .git/18:18
→ kwmiebach joined18:18
LindsayMac She's using a GUI for git called Source Tree.18:18
An_Ony_Moose: ok i will relay the message. Thank you18:19
sm0ke it just has a commit which i just made followed by commit which is did 1 month back..in betweenthe upstream repo has many commits which got merged into my latest commit18:19
→ frgtn joined18:19
sm0ke wtf why does git work that way why doesnt fetching upstream fetches the commits seperately instead of merging them in latest commit18:19
skorgon fetch never merges18:20
An_Ony_Moose ^18:20
← v0n left18:20
skorgon and i doubt that git merges anything "into" your commits18:20
← GODDOG left18:20
→ youlysses joined18:20
skorgon if you pull it'll merge into your current branch, but well, that's what you ask for when you pull18:20
→ Prazanna joined18:21
→ v0n joined18:21
← Kronuz left18:21
skorgon sm0ke: you should be able to find the HEAD revision your branch was at before the merge you try to undo. find that commit and create a new branch pointing at that commit. from there you should be able to go whatever way you deem to be right. rebase, merge, neither or something else18:22
← offby1 left18:22
→ offby1 joined18:23
offby1 !gka18:23
gitinfo For a better way to view the reflog, try: gka() { gitk --all $(git log -g --format="%h" -50) "$@"; }; gka18:23
offby1 sm0ke: that's for you ^^18:23
easy way to find out the previous commit18:23
← arietis left18:25
→ mklappstuhl joined18:26
→ Goplat joined18:26
sm0ke offby1: AHA nice18:26
→ Kronuz joined18:26
← Tuna-Fish left18:26
→ Baluse joined18:27
→ Tuna-Fish joined18:27
sm0ke now i see a subtree which has all the changes in between my two commits18:27
← koda left18:27
sm0ke what do i do now?18:27
← mklappstuhl left18:27
← d3c left18:27
→ dreiss joined18:28
← travisstaton left18:28
→ arietis joined18:29
→ cmr joined18:29
sm0ke reading upon advice you gave how do i create a new brach pointing at a particular commit18:30
← v0n left18:30
sm0ke i can see the SHA18:30
← CKyle left18:30
skorgon git branch <name> <commit>18:30
or if you want to check it out immediately 'git checkout -b <name> <commit>'18:31
→ jruiz joined18:32
→ sq-one joined18:32
→ v0n joined18:32
← v0n left18:32
→ v0n joined18:33
→ bwreilly joined18:33
→ chlue joined18:34
← frgtn left18:35
→ mklappstuhl joined18:35
→ frgtn joined18:35
LindsayMac love it when you enter a hackathon, someone says they have experience with git, and then spends 4 hours asking about git18:35
→ h0bbit_ joined18:36
cbreak what's a hackathon?18:36
LindsayMac 'hackathon is a set time where teams get together and build SOMETHING that will benefit the world/planet/environment/people18:36
sm0ke crap i need to stash changes i dont want a resolve common conflicts .. be back in 5 mins18:36
LindsayMac then at the end the ideas/projects are judged and a winner is selected18:36
cbreak sounds weird.18:36
← mklappstuhl left18:36
LindsayMac surprised so many people haven't heard of or done a hackathon18:36
cbreak good luck.18:36
LindsayMac yea well.. the person with all of the style files can't give them to us18:37
cmr I'm trying to get a commit timestamp for a given commit. `git show --format=%ct -s $sha` works for some things, but also dumps lots of stuff I don't care about when the sha is, for example, a tag. Is there a plumbing command that can help me here?18:37
→ CKyle joined18:37
cmr LindsayMac: that's only one kind of hackathon18:37
LindsayMac its the common definition18:38
grawity cmr: how about `git log -1 --format=%ct $sha`18:38
cbreak show needs a --no-diff option.18:38
sm0ke skorgon: nice.. i created a new branch pointing to just before my change with all the upstream changes let me push that...how do i now get my last commit from previous branch now18:38
cmr cbreak: that's what -s is18:39
grawity: that does it, thanks! what does -1 do? I don't see it in the man page18:39
grawity cmr: a shortcut for -n 118:39
cmr Ah18:39
skorgon sm0ke: i'm not 100% sure i have the correct picture of your branche's histories, but i guess you want to cherry-pick that one missing commit into your new branch18:40
← frgtn left18:40
← RaceCondition left18:40
→ Pyrogerg joined18:40
→ tatsuya_o joined18:40
→ mklappstuhl joined18:41
sm0ke OK i fu**ed everything up now18:41
← mklappstuhl left18:42
← tatsuya_o left18:42
sm0ke i did a git push from my new branch .. doesnt github creates a new branch for me?18:42
→ Sam_OTnosc joined18:42
← freestyl3r left18:43
→ tatsuya_o joined18:43
← Balliad left18:43
skorgon might depend on your push settings and used syntax.18:43
→ freestyl3r joined18:43
sm0ke i just did a "git push"18:43
→ mklappst_ joined18:44
← DonRichie left18:45
← rouf left18:45
→ Balliad joined18:46
cbreak git push doesn't create new branches18:46
(unless you explicitly configure it that way)18:46
grawity I think the new push.default does18:46
cbreak tell git to push a specific branch if you want that18:46
← JEEB left18:46
→ rouf joined18:46
→ frgtn joined18:46
sm0ke ok i am in deep shit now18:46
cbreak grawity: it configures upstream to some non-existing branch?18:47
← LuxuryMode left18:47
← tatsuya_o left18:47
→ yhager|away joined18:47
← Advocation left18:48
← Sam_OTnosc left18:49
sm0ke ok so now i have my stupid commit on my forked repo...just need to find a way to delete this commit i guess18:49
← staafl left18:49
← rouf left18:50
→ staafl joined18:50
skorgon sm0ke: sort things out locally in some branch. and once you're happy with that branch (force) push it to publish it18:50
← dreiss left18:50
← jruiz left18:51
sm0ke skorgon: yea trying that18:51
← rat_thing left18:51
← Kronuz left18:52
→ threesome joined18:53
sm0ke skorgon: got it 1) create a branch pointing to latest commit on upstream branch 2) push ( this time without fail by) git push origin dev218:55
← shoshin left18:55
→ Fernando1asso joined18:55
sm0ke everything seems ok .. same no. of commits in my fork and upstream development branch18:56
→ dreiss joined18:56
→ khushildep joined18:56
sm0ke now how do i cherry pick my transction from another branch ( orginally on which i was working) ?18:56
→ JEEB joined18:56
sm0ke my commit*18:56
wow i dont know shit about git18:57
hey that rhymed18:57
← Balliad left18:58
→ ehmatthes joined18:59
← FernandoBasso left18:59
← frgtn left18:59
sm0ke i guess i just need the commit hash?19:00
→ frgtn joined19:00
← yhager|away left19:00
skorgon right, just git cherry-pick <commit>19:00
← hoschi left19:00
← shuaib left19:01
→ Balliad joined19:01
sm0ke thanks i think i got it just failed .. but thats due to confilicting file19:02
← vkz left19:02
← frgtn left19:02
→ jnavila joined19:04
← mklappst_ left19:04
skorgon just resolve the conflict(s), add the change and run git cherry-pick --continue (IIRC, git status tells you the same)19:04
sm0ke oh crap..is that necessary i just manually added all files for commit and pushed the commit19:06
everything looks fine though19:06
→ kerrick joined19:06
skorgon run git status. it might still think you're cherry-picking19:06
← Dougie187 left19:07
sm0ke no i just see untracked files19:07
→ shoshin joined19:08
← devoldmx left19:09
skorgon then you should be fine. i hope with "i just manually added all files for commit", you mean, that you resolved the merge conflicts.19:09
→ devoldmx joined19:09
→ Davey joined19:10
sm0ke skorgon: yes i did..i remove the weird markers >>>> and ==== manually19:10
skorgon: thanks much!19:10
skorgon okay19:10
np19:10
← RobW_ left19:10
→ harobed joined19:12
← rudak left19:12
→ rudak1 joined19:12
← thierryp_ left19:12
← Davey left19:12
→ wizard_A joined19:12
sm0ke just need to clean up everything but may be in morning..got to sleep now..how sometimes a 5 minute works turns into 1 hour19:14
wizard_A when i issued a pull command it asks for username and password, i entered both but after this there is a great pause in the terminal.19:14
→ thierryp_ joined19:14
sm0ke learned a few thing though19:14
have a good one.19:14
← sm0ke left19:15
skorgon sm0ke: well, at least you just had to do soem git magic and not rewrite your code19:15
← nemesit|znc left19:15
→ yhager|away joined19:15
→ rat_thing joined19:15
→ beastd joined19:17
wizard_A i deleted my master remote is that why it is not able to merge.19:17
→ Mars` joined19:17
→ nemesit|znc joined19:19
→ frgtn joined19:19
← zarubin left19:19
→ mikef joined19:20
→ shuaib joined19:20
← shuaib left19:20
→ Gamecubic joined19:20
→ ewokchewy joined19:22
→ vigs joined19:23
→ tatsuya_o joined19:24
→ Dunedan joined19:25
→ |Freeman| joined19:25
← Dunedan left19:26
← rat_thing left19:26
← ewokchewy left19:27
← Jackneill left19:28
→ Saurabh_ joined19:31
→ mklappstuhl joined19:31
← axs left19:31
← yhager|away left19:33
→ p3lim joined19:33
← CheckDavid left19:34
← Fernando1asso left19:34
→ zarubin joined19:36
← joshsmith left19:36
← arietis left19:36
→ cirenyc joined19:37
→ idoru joined19:37
← raonyguimaraes left19:38
→ kamikaz3 joined19:39
→ arietis joined19:40
← mikef left19:41
← wizard_A left19:41
→ mklappst_ joined19:43
→ devoldmx3 joined19:44
← mklappstuhl left19:45
← devoldmx left19:45
→ rat_thing joined19:45
→ therealGent joined19:45
→ Orbitrix joined19:46
← Orbitrix left19:46
→ Orbitrix joined19:46
← asQuirreL left19:47
→ mikef joined19:47
← rat_thing left19:47
→ cff joined19:48
→ jchen joined19:48
jchen if I need to move the path where a submodule currently lives, how would I go about doing so?19:48
← a-l-e left19:49
jchen do it deinit the submodule then re-add it to the new location?19:49
← robbyoconnor left19:50
← mklappst_ left19:52
← cirenyc left19:52
thiago git mv19:53
→ Sendoushi joined19:53
thiago or do you mean the repository URL?19:53
→ gondalier_ joined19:53
jchen no, the actual dir location19:53
git mv doesn't work: fatal: source directory is empty19:53
← vurma left19:54
thiago is it already initialised?19:54
jchen yes19:54
← Gamecubic left19:54
thiago so the directory is not empty, right?19:55
→ robbyoconnor joined19:55
OMGOMG !quiz hint19:55
← Sendoushi left19:55
jchen right, there's stuff in there19:55
gitinfo Error: I'm not a quiz master19:55
→ Shaladdle joined19:56
→ Sendoushi joined19:56
thiago jchen: then try to mv manually, git rm the old one, use the normal git submodule stuff to add the new path19:56
← ciampix left19:56
→ framling joined19:56
jchen tahts what i'm trying to avoid, cuz that seems really jank19:56
unless that's how it's supposed to be done?19:56
→ snearch joined19:57
→ jalama joined19:57
OMGOMG gitinfo: sorry19:58
jchen also fwiw, i've been doing rm, then git mv all the stuffs and then trying to re-initialize19:58
but it doesn't work as expected because there are still references to the old path in .gitmodules19:58
← robbyoconnor left20:00
→ pantsman joined20:01
← pantsman left20:01
→ pantsman joined20:01
← arietis left20:02
← _Harm_B_ left20:03
→ _Harm_B_ joined20:03
→ CheckDavid joined20:04
← shoshin left20:04
→ Risto- joined20:04
→ Devastator_ joined20:05
← mikef left20:05
← Devastator_ left20:05
→ Devastator_ joined20:05
→ shoshin joined20:05
← milestone left20:05
← mikepack left20:06
← Devastator left20:06
Devastator_ → Devastator20:07
→ mikepack joined20:07
← Mars` left20:07
→ Mars` joined20:07
← chairabanta left20:08
→ Gamecubic joined20:08
→ milestone joined20:08
← TomyLobo left20:10
← jalama left20:10
← mikepack left20:11
← Mars` left20:12
← milestone left20:12
→ devoldmx joined20:13
→ cirenyc joined20:13
→ LuxuryMode joined20:13
← beastd left20:14
← gondalier_ left20:14
← chisler left20:14
← devoldmx left20:15
← bitkiller left20:15
← devoldmx3 left20:15
→ LBV_User joined20:15
→ devoldmx joined20:16
← CKyle left20:16
→ gondalier joined20:18
→ rpg joined20:19
← roxlu left20:20
← arand left20:20
→ robbyoconnor joined20:21
→ AWinterman joined20:21
← Risto- left20:22
← gondalier left20:22
→ RaceCondition joined20:22
← h0bbit_ left20:23
→ Mars` joined20:25
← Destos left20:25
→ gondalier joined20:26
← staafl left20:26
← thomas_b left20:28
← gondalier left20:29
← cmpera left20:30
← devoldmx left20:30
→ devoldmx joined20:30
→ ahri joined20:32
→ devoldmx3 joined20:32
ahri → Guest3753920:33
→ lotus joined20:34
← shoshin left20:35
← longbeach left20:35
← devoldmx left20:35
← sq-one left20:36
← Spyro left20:37
← devoldmx3 left20:37
← Prazanna left20:38
→ Destos joined20:40
← peroht left20:40
← Destos left20:41
← Balliad left20:45
→ PaulCapestany joined20:45
→ Destos joined20:45
← Destos left20:45
→ Spyro joined20:46
← daanavitch left20:46
→ theRealGentleman joined20:46
→ thomas_b joined20:46
→ Balliad joined20:47
→ valgrind joined20:47
← kongthap left20:47
← therealGent left20:47
← AWinterman left20:47
→ AWinterman joined20:48
← Sendoushi left20:51
← Balliad left20:51
← cirenyc left20:52
← AWinterman left20:52
← petey left20:53
→ Prazanna joined20:53
→ petey joined20:53
→ koda joined20:53
→ linukso joined20:53
→ Balliad joined20:53
linukso It's git transfer of sources reliable?20:54
In other words, could I trust the downloaded source?20:54
← danicampa901 left20:54
thiago yes, it's totally reliable20:55
it's cryptographically secure20:55
cmn if you trust the source of the data20:55
skorgon linukso: depends on how much you trust the server and the transport channel20:55
← _iron left20:55
thiago you don't need to trust the server or the transport channel20:55
skorgon you can sign tags though20:55
thiago Git can verify that the transfer was correct20:55
skorgon thiago: but not that it was not tempered with20:56
thiago that's implied20:56
you need something to compare *with*20:56
skorgon well, that why you can signed tags, right? so you don't have to trust anybody but the signing key20:57
thiago it's like connecting to an https server that has a self-signed certificate: if you accept it, you know that you're connected to the server. You just don't know if it's the right server.20:57
that's one way20:57
I could just tell you the SHA-1. If you trust me or if you can get the SHA-1 from another source, it's enough.20:57
skorgon right20:57
← petey left20:57
thiago the public signature is just a way to confirm that the SHA-1 was the one you wanted.20:57
→ ping-pong joined20:57
skorgon that's why i said, if you trust the origin and transport channel, you're fine20:57
thiago you transfer the trust to the public key. Do you trust that key?20:58
→ rat_thing joined20:58
← LuxuryMode left20:58
iveqy linukso: if you can verify that the tip of the branch you're interested in has the same sha1 as the one you trust, then you can trust all history and all source20:58
thiago if you trust the origin and the transport channel, you don't need to verify SHA-1 or tags20:58
you know it's right20:58
iveqy thiago: yes you do, otherwise you don't check for harddisc errors20:59
thiago that's something completely different. That's corruption, not trust.20:59
skorgon thiago: faking a git identity is easier than a signed tag20:59
thiago skorgon: that assumes the origin and transport channel couldn't be trusted.20:59
iveqy with git, you never need to trust the transport channel nor the origin. The transport channel is harmless20:59
skorgon thiago: right21:00
← Misan left21:00
iveqy thiago: I would say that's trust too21:00
thiago what iveqy said now: you don't need to trust the transport channel or the origin21:00
iveqy corruption is just an other evil21:00
thiago you need to get some out-of-band data to compare with21:00
examples: the SHA-1 written in a piece of paper, hand-delivered by someone you know and trust21:00
→ noname001 joined21:00
→ cirenyc joined21:00
thiago the public key of someone you trust, who signed a tag21:01
← _nova left21:01
thiago 10000 copies of the repository in multiple servers, reporting the same SHA-1, plus posts on the mailing lists21:01
iveqy however, most people don't only need to trust their source, they also need to be sure that noone else can access their source. In that case, the answer is a bit more complicated21:03
← dfanjul left21:03
← rat_thing left21:03
thiago trust always comes via an out-of-band piece of information21:03
skorgon iveqy: why? a signed tag should sovle exactly that? if somebody alters your code on the server the signature check won't pass21:04
thiago right21:04
but that's because the SHA-1 of the tag wouldn't match21:04
because the SHA-1 of the commit wouldn't match. If you can verify the SHA-1 of the commit, you don't need the tag.21:04
iveqy skorgon: yes, but that doesn't prevent someone else from stealing your code, just insert evil code inside your project21:04
skorgon thiago: but i as downstream user probably don't know the correct sha121:05
thiago the tag helps because you can use the signature embedded in it as a verification21:05
exactly21:05
→ dfanjul joined21:05
→ Caritade joined21:05
skorgon iveqy: i don't follow. if you always push signed tags and tell your users to verify the tags before building you're good. even if something evil is in your repo, you would recognize it by a failing signature check and throw it away21:07
← threesome left21:07
← penguin42 left21:07
→ yhager|away joined21:08
→ Asenar joined21:08
← geak left21:08
→ HerrK joined21:08
← vigs left21:09
→ geak joined21:09
→ milestone joined21:09
linukso holy sh...21:09
→ ErKa joined21:09
← Xethron left21:09
← graphitemaster left21:09
linukso so, what's the conclusion? I'm asking for git own transport21:10
it implements any TLS? it have WoT + GPG?21:10
→ cads joined21:10
skorgon i guess the question is, what are you concerned about?21:10
how do you clone?21:10
via http or git, it's insecure21:10
ssh uses ssh, so secure21:10
https uses ssl and is encrypted as well21:11
→ graphitemaster joined21:11
← cff left21:11
thiago the git protocol is unauthenticated and unencrypted. Git will take care of verifying the integrity of the data by itself.21:12
the git protocol will prevent corruption problems21:12
← noname001 left21:13
← PeterBooker left21:13
→ Sendoushi joined21:14
← irqq left21:14
linukso thiago, I'm saying about security to21:15
← tvw left21:15
← bwreilly left21:15
linukso well, we wait for some tls implementation in git protocol...21:15
cmn that won't happen21:16
linukso cmn, could you see the future?21:16
← cirenyc left21:17
cmn I don't see adding enormous complexity so something that doesn't need it when there are at least two transports that already provide anything you could ask of it21:17
→ petey joined21:17
linukso or whatever that satisfies integrity and authenticity21:17
skorgon there is transport via ssh and https... what esle do you need?21:18
linukso skorgon, just it21:18
most of the servers don't provide it21:18
just git protocol21:18
http://git.savannah.gnu.org/cgit/hurd/hurd.git/21:18
← gitpicker left21:18
← petey left21:18
linukso ssh, but do I have access?21:18
cmn those are not interested then21:19
skorgon why would you want encrypted transport to an FOSS project?21:19
that doesn't make sense21:19
thiago use https for anonymous, encrypted support21:19
→ petey joined21:19
← koda left21:19
cmn why would they deploy this additional layer on top of the tcp transport when they won't for one that does have it?21:19
← khushildep left21:19
skorgon it's open source and will be published on the web anyway, why would you go through the overhead of encryption for that?21:19
linukso skorgon, sure21:19
→ osmosis joined21:19
linukso cmn, yes I'm asking that, https21:19
thiago if you need authentication, like we discussed before, you need a piece of out-of-band information confirming you've reached the right server21:19
linukso skorgon, why not?21:20
cmn asking what?21:20
skorgon and github and similar services usually use ssh21:20
thiago that usually means the CA signing the server's certificate21:20
← rudak1 left21:20
skorgon and actually most other services for push as well, since ssh provides auth21:20
linukso I just need https21:20
cmn use that then21:20
thiago linukso: then use https21:20
cmn why add TLS to git://?21:20
linukso but most of the servers don't provide it21:20
skorgon linukso: because it costs resources for no benefit21:20
linukso cmn, not TLS21:20
thiago linukso: what should it use, if not TLS?21:21
linukso <linukso> or whatever that satisfies integrity and authenticity21:21
thiago Git itself satisifes integrity21:21
authentication requires a piece of out-of-band information21:21
cmn linukso: yes, you just did ask for that21:21
cbreak linukso: if you need https, then use https21:21
but ssh is better. :)21:21
linukso cbreak, why it should be better?21:21
thiago TLS solves that by using the CA ceritificate-signing method21:21
cbreak linukso: it's not that it should be better21:21
linukso: it is that it is better21:22
linukso cbreak, I'm asking why not saying21:22
skorgon, man, I'll give you a project with 1 gSLOC by git21:22
cbreak ssh is better than https because it is a faster and more streamlined protocol than https21:22
linukso skorgon, open source, just with http, you'll run it?21:22
cbreak it is easier to set up for git21:22
it has way stronger client authentication by default21:23
thiago cbreak: do you mean git-over-ssh is faster and more streamlined than git-over-https?21:23
linukso cbreak, yeah that's true21:23
cbreak, by default21:23
cbreak, but21:23
← petey left21:23
skorgon linukso: i don't get what you want. for authenticated access use git via ssh.21:23
cbreak thiago: yes. http is still a pull based protocol, and the pure existence of dumb http pulls down the average streamlining and speed21:23
linukso cbreak, SSH needs WoT, TLS uses centralized trustworthy21:23
cbreak linukso: no.21:23
linukso cbreak, yes21:23
thiago cbreak: how about the smart http?21:23
cbreak linukso: use your brain a bit :(21:24
linukso cbreak, always21:24
cbreak SSH doesn't even have a WoT21:24
skorgon linukso: which has proven so secure...21:24
linukso cbreak, what the f...21:24
thiago SSH doesn't have a WoT21:24
you are supposed to get the server's key or fingerprint out-of-band21:24
linukso cbreak, I have a WoT for ssh, the place that's trustworthy were I list my SSH keys21:25
that's a verification for my assinatures in SSH21:25
← knigitz left21:25
linukso that's SSH WoT21:25
→ LuxuryMode joined21:25
cbreak it seems you don't know what a web of trust is.21:25
← Nicholas_Rishel left21:25
cbreak honestly, https has more web of trust like features than ssh will ever have21:26
patagonicus There's a tool that uses GPG for SSH auth, if I remember correctly. Not directly WoT, but you can just check if you trust a key and if so add it to authorized_keys. Same for host keys.21:26
cbreak but with https you have to trust designated CAs21:26
→ Nicholas_Rishel joined21:26
linukso cbreak, and with SSH?21:26
cbreak with ssh there's no one to trust but yourself, because YOU have to verify the keys21:26
linukso cbreak, do we have CA's?21:26
cbreak, SURE21:26
cbreak, that's why SSH needs WoT21:26
cbreak but it doesn't have one.21:26
linukso to verify signatures21:26
cbreak, I use one21:27
thiago that's outside the protocl21:27
linukso thiago, SURE IT IS!21:27
cbreak linukso: what you use is not ssh21:27
thiago it's the "out-of-band" data that we discussed21:27
linukso GPG WoT is TO!21:27
→ nathancahill joined21:27
cbreak linukso: because ssh does not have a web of trust21:27
linukso: you should REALLY read up on the basics of the technologies you use21:27
linukso cbreak, and what I said? do you have comprehension problems?21:27
I said SSH needs21:27
not it haves21:28
thiago it will hardly get one now21:28
cbreak linukso: it doesn't need21:28
thiago ssh has been like that for 15 years21:28
linukso sigh...21:28
cbreak, it needs21:28
thiago anyway, that's totally off-topic21:28
cbreak linukso: if you want to continue to blabber meaningless garbage, go somewhere else.21:28
← thierryp_ left21:28
thiago git uses ssh assuming that you've done the necessary steps to verify that it's secure21:28
linukso cbreak, who's doing that is you, man21:28
thiago git also uses https assuming you've done the same21:29
skorgon let's go back a few minutes. i think linukso is right, he should wait for git:// implementing tls21:29
linukso skorgon, not TLS21:29
→ thierryp joined21:29
thiago if your use of ssh or https isn't correct, then your weakest link is there21:29
linukso skorgon, could be other thing that gives us authencity and integrity21:29
thiago git doesn't need one more authenticated, encrypted protocol. It already has two.21:29
linukso but git protocol is perfect for git transfers, but it's insecure!21:29
thiago explain why those two aren't enough for your needs21:30
cbreak git:// is meant to be insecure.21:30
← chlue left21:30
linukso cbreak, meant?21:30
skorgon linukso: this is going in circles. git via ssh gives you exactly what you want. use it or don't, i don't give a shit21:30
cbreak intended to be.21:30
linukso thiago, my needs?21:30
thiago linukso: your needs21:30
linukso thiago, well, git protocol is perfect for git transfers21:31
iveqy git as built in integrity (the sha1) for all protocolls. Git has three authenticated protocolls (http, https, ssh) and git has two encrypted protocolls (https, ssh). What's the fuzz about? Just use whatever you need...21:31
← osmosis left21:31
thiago iveqy: http is authenticated?21:31
iveqy thiago: yes21:31
thiago iveqy: what do you mean?21:31
cmn linukso: stop repeating that; we know, if you want to put encryption around it, use the ssh or https transports21:31
cbreak authentication without encryption is worthless21:31
linukso cmn, I do so21:31
cbreak because it is suceptible to MitM attacks21:31
iveqy thiago: http://en.wikipedia.org/wiki/Basic_access_authentication21:32
cmn then stop complaining21:32
thiago iveqy: like cbreak said: authentication without encryption is worthless21:32
iveqy cbreak: yes, but http still has it21:32
thiago iveqy: it's also client authenticating to the server. I thought we were discussing server authentication.21:32
→ Kronuz joined21:32
iveqy hey, I didn't designed it:(21:32
thiago iveqy: how do you verify that you reached the right server?21:32
linukso cbreak, what bullshit, we could protect against MiTM without encryption21:32
thiago linukso: no, we can't21:32
iveqy thiago: oh, that you can't do with http as far as I know.21:32
linukso without encrypting the data I meant21:33
thiago, yes we cam21:33
← thierryp left21:33
cbreak linukso: no trolling please.21:33
thiago linukso: you can verify the integrity of the data that you received. But you can't tell whether the data you received is what it was supposed to be.21:33
← alezost left21:33
thiago linukso: what use is to verify that malicious data isn't corrupt?21:33
linukso thiago, we could sign the data21:33
thiago, signing isn't encryt the data21:34
thiago linukso: that's the out-of-band information again21:34
linukso it uses encryption21:34
thiago right21:34
it uses encryption. Public key encryption, to be precise.21:34
linukso thiago, I did, if you read what I said21:34
<linukso> without encrypting the data I meant21:34
thiago linukso: ok, that is true21:35
linukso so is the life, can't download hurd with secure sources...21:35
thiago linukso: you've just described the git:// protocol21:35
linukso thiago, but it just uses SHA doesn't?21:35
thiago linukso: don't trust the origin, don't trust the channel. All you need is to verify the integrity of the data and public key encryption to verify a tag.21:35
cbreak linukso: ever heard of signed tags? signed commits?21:35
linukso thiago, I sincerely want to21:36
thiago for example, I can download the kernel source code from kernel.org or from h4x0r.org. It doesn't matter.21:36
all I need is Linus's public key so I can verify that the tag he signed passes the check.21:36
cbreak did you know kernel.org got hacked?21:36
they still didn't find out how21:37
linukso cbreak, but not the keys21:37
cbreak so they still didn't fix the hole21:37
thiago cbreak: I think they did21:37
cbreak thiago: they did finally? neat :)21:37
linukso thiago, I sincerely want to trust te channel, I want to do not waste CPU with encryption but I have to21:37
cbreak either way, the git history is still safe21:37
thiago linukso: why do you want to trust the channel?21:37
cbreak why? because it's signed.21:37
thiago linukso: if you want to trust the channel, you want encryption.21:37
iveqy linukso: trusting the channel is nonsense for an opensource project21:37
linukso thiago, not21:37
thiago, no21:37
iveqy, it is21:37
← Balliad left21:38
linukso for now it is21:38
thiago linukso: sorry, you can't have it both ways. If you want to trust the channel, you need encryption.21:38
← snearch left21:38
cbreak if you want to trust the channel, then you need to put down the ethernet cable yourself.21:38
thiago linukso: if you want to transmit without encryption, you don't trust the channel. But you can verify whether the data was corrupted.21:38
linukso thiago, no man, what I want is trust the channel, what I want is to anyone to hijack my connection21:38
thiago linukso: we're having communication problems21:39
linukso no one with malicious purposes, what I meant21:39
thiago linukso: once again: trust channel = encryption21:39
→ staafl joined21:39
← Guest37539 left21:39
thiago trust source = authentication21:39
linukso thiago, man, did you ever heard about quantic cryptography?21:40
ChanServ set mode: +o21:40
thiago linukso: how is that relevant?21:40
linukso not cryptography, it doesn't encrypt the data21:40
cbreak linukso: honestly...21:40
thiago cbreak: don't take harsh measures21:40
linukso but it is secure21:40
thiago linukso: is it in use now?21:40
← Kakera left21:40
cbreak thiago: you think he's seriously not trolling?21:40
thiago cbreak: up until the point about quantum encryption now, no21:41
linukso thiago, you said trust = cryptography21:41
thiago linukso: no21:41
linukso: I said trust channel = encrypted channel21:41
linukso thiago, yes that what I meant21:41
cbreak linukso: get a clue.21:41
cbreak set mode: -o21:41
← victorbjelkholm left21:41
thiago linukso: to the best of our knowledge of computers and problems, encryption is secure21:41
→ Balliad joined21:41
linukso thiago, yes but21:41
thiago linukso: no "but"21:41
linukso: stop at "yes"21:41
→ NetJunky joined21:42
linukso thiago, what I remember for quantum encryption21:42
thiago linukso: stop21:42
linukso: we agreed that encryption is secure, to the best of our knowledge21:42
NetJunky Hi everyone!21:42
linukso that it just don't encrypt the data man, WHY can't you fucking listen to me?21:42
thiago linukso: because you're not making sense21:42
linukso thiago, wait a second21:42
thiago linukso: and do NOT use swear words. One more and you will get kicked out of this channel.21:43
cbreak linukso: quantum cryptography is for key exchange, not data transmission. It requires a dedicated link, no switches, no routers, no anything in between. So it is completely impractical for almost everything.21:43
← VictorCL left21:43
← linukso left21:43
thiago oh well21:43
← nathancahill left21:43
thiago cbreak: I think that was a completely clueless person who looked at how Git works and though, "I can revolutionise security!"21:44
NetJunky Guys, do I understand correctly, that git cherry-pick is something like Ctrl+C and Ctrl+V one commit from one branch to another? Except, that on branch where it is pasted it is received as entirely new commit?21:44
thiago obviously seriously misled21:44
cbreak thiago: clueless, yes.21:44
thiago NetJunky: yes21:44
NetJunky: it means "take diff, apply diff"21:44
NetJunky @thiago: so this would mean, that it clusters the history.21:45
cbreak NetJunky: it copies the changes and the commit message, and creates a new commit with those and the current commit as parent.21:45
thiago NetJunky: correct21:45
← leeN left21:45
NetJunky @cbreak, @thiago: thank you21:45
cbreak the only relation with the original is the patch id21:45
NetJunky: see man git-patch-id21:45
gitinfo NetJunky: the git-patch-id manpage is available at http://jk.gs/git-patch-id.html21:45
→ linukso joined21:45
thiago NetJunky: hint: this is IRC, don't use @. The nickname is enough.21:45
cbreak @ means administrator :D21:46
at least in my client21:46
thiago for everyone, and technically "op"21:46
NetJunky Okey. Thank you.21:46
thiago admin is something else21:46
cbreak yes.21:46
linukso back, look, why quantum channel ins't for data trasnfer?21:46
← Kronuz left21:46
thiago linukso: do you have one available?21:46
linukso we could exchange keys, it would be better21:46
NetJunky cbreak: will take look at it. Thank you.21:46
linukso thiago, well, didn't you said trust channel = encryption21:46
cbreak linukso: the same reason why DH is not for data exchange21:47
thiago linukso: we're talking about real life.21:47
linukso: do you have ANY quantum channel?21:47
linukso thiago, doesn't matter21:47
thiago linukso: sure it does21:47
cbreak linukso: because you CAN NOT control the result21:47
thiago linukso: we're talking about real life. Are there any quantum channels deployed on the Internet that we could use?21:47
cbreak the longest quantum link I've heard of was a few km tops21:48
linukso cbreak, but we could have authenticity and integrity not confidentiality21:48
cbreak it was a line-of-sight communication21:48
linukso: no.21:48
thiago linukso: totally and completely irrelevant21:48
cbreak linukso: it is for key exchange.21:48
linukso: what do you think that key is used for?21:48
thiago linukso: quantum channels are not in use today. Therefore, they are irrelevant.21:48
linukso cbreak, do you are the king to say for what we should use quantum channels?21:48
→ bwreilly joined21:48
linukso so trust channel != encryption21:48
that ends here21:49
cbreak linukso: stop being a complete moron and think21:49
thiago cbreak: and stop insulting him21:49
linukso cbreak, I know man I know21:49
NetJunky Guys quick Q about workflow as well.21:49
cbreak quantum channels are NOT trusted21:49
← jnavila left21:49
NetJunky Does anyone one of you know PHP at any level?21:49
offby1 shifts uncomfortably21:49
cbreak NetJunky: enough to make fun of it.21:49
thiago cbreak: please don't enter the discussion about what a quantum channel does or does not do.21:49
linukso cbreak, yeah stop that could you use your brain for the discussion? and not for the offenses21:49
thiago linukso: quantum channels are NOT in use. Period.21:49
cbreak linukso: if you can't handle the truth, don't ask.21:50
thiago linukso: it doesn't matter whether they are secure or not. They are irrelevant.21:50
patagonicus NetJunky: Yes, at the "I wrote a dozen lines in it once and never want to use it again" level. But just ask, you have higher chances of getting an answer.21:50
thiago linukso: so please tell us what your problem is21:50
linukso quantum channels are trustworthy21:50
thiago linukso: quantum channels are irrelevant21:50
cbreak they are not...21:50
thiago linukso: stop talking about quantum channels.21:50
linukso they are, we could say if the data is trustworthy or not21:51
← SamB left21:51
thiago linukso: they are irrelevant. They are not in use.21:51
→ Kronuz joined21:51
linukso thiago, Yeah I know, but21:51
← zarubin left21:51
thiago linukso: no "but"21:51
NetJunky cbreak: :) Well I have a project, that supports older versions like 5.3 and lower and newer versions. Newer version is at master branch and older at branch A. Now if I have a feature branch, that can be applied to both long-term branches, on what branch should I base it?21:51
offby1 I wish _everyone_ would quit talking about quantum channels; if I were to ignore you all, there'd be few intelligent people left21:51
thiago linukso: since they are not in use, we can't use them.21:51
cbreak NetJunky: on the merge base of the two branches?21:52
thiago linukso: the question is how we can achieve secure transmissions over channels we *do* have. That's, to the best of our knowledge, encryption.21:52
← youlysses left21:52
offby1 NetJunky: have you created this feature branch yet?21:52
linukso thiago, to make affirmatives we needs to know if it's true21:52
offby1 also, what cbreak said :)21:52
NetJunky cbreak: didn't understand.21:52
offby1: not yet.21:52
← ngladitz left21:52
cbreak NetJunky: the merge base is the last commit those two branches have in common21:53
NetJunky: man git-merge-base21:53
gitinfo NetJunky: the git-merge-base manpage is available at http://jk.gs/git-merge-base.html21:53
NetJunky cbreak: you mean the point when two long term branches diverged?21:53
thiago linukso: do you want to continue this discussion based on the premise "encryption is not secure" ?21:53
cbreak yes.21:53
linukso thiago, did I said that?21:53
thiago linukso: I am asking if that's what you mean21:53
← harobed left21:53
linukso thiago, no, its not what I meant, I meant it ins't just encrypted-based channel that's secure21:54
thiago linukso: I agree with you21:54
NetJunky cbreak: I see. So I should base feature branch on commit and not some specific branch, that already exists.21:54
thiago linukso: there are many secure channels21:54
offby1 NetJunky: that comment makes no sense21:54
thiago linukso: how many of them are widely available for people on the Internet?21:54
→ cagedwisdom joined21:54
linukso thiago, so why did you said trust channel = encryption?21:54
NetJunky cbreak: how will it be applied? Can't imagene, sorry.21:55
→ DrShoggoth joined21:55
← Mars` left21:55
cbreak NetJunky: make a new branch from that commit21:55
→ gondalier joined21:55
offby1 NetJunky: I think cbreak means that you should base your branch on the newest commit that is common to your existing branches.21:55
cbreak done.21:55
→ petey joined21:55
thiago linukso: because we were discussing transmissions over the internet. I don't know any other secure channel there.21:55
← mribeirodantas left21:55
thiago linukso: do you?21:55
→ Mars` joined21:55
NetJunky cbreak: I see. Thank you.21:55
offby1: understand.21:55
linukso thiago, I don't know, but it not means it ins't there21:56
← yhager|away left21:56
thiago linukso: I agree. There may be others.21:56
linukso: but since neither you nor I know about them, how would we design software to use them?21:56
linukso thiago, and I'm not discussing reliability21:57
← lotus left21:57
cbreak nfs works that way.21:57
thiago linukso: nor am I21:57
→ lotus joined21:57
thiago linukso: I'm discussing practicality21:57
linukso I'm discussing security21:57
cbreak it assumes you're on a lan and control all computers.21:57
thiago linukso: so am I21:57
jon_ does anyone use Atlassian Stash? We're considering switching to it and would love some feedback21:57
linukso thiago, practically quantum encryption is secure21:57
→ shinnya joined21:57
thiago linukso: but we already established that it's not practical for use.21:57
linukso: so that's out21:57
linukso: what other secure channels do you know?21:58
linukso thiago, let me see21:58
patagonicus Also, what do you mean by "secure". That isn't a simple question, there are lots of different attack models.21:58
← RaceCondition left21:58
← gondalier left21:58
linukso thiago, pre-know encryption methods21:58
→ SamB joined21:58
linukso thiago, or keys, or whatever21:58
thiago linukso: one-time pads?21:58
linukso thiago, I think one-time pads is dependent of ..21:59
← mishok13 left21:59
thiago well, you're describing encryption now21:59
linukso thiago, yes I'm21:59
thiago do you know any secure channel that doesn't rely on encryption?21:59
linukso thiago, without encryption?21:59
→ epicowl joined21:59
ggherdov Hello. I cloned a repo getting changesets A <-- B <-- C. I added my cset D on top, thus now I have A <-- B <-- C <-- D. In the meantime, in upstream they committed E on top of C, and they have A <-- B <-- C <-- E. I would like to pull E from upstream, w/o mergind with D; I'd rather like to have two heads on my local repo. Then, my next step would be to21:59
rebase D on top of E, since I want my cset D to be the latest, in my local repo.21:59
How is that done?21:59
← Mars` left22:00
thiago ggherdov: git rebase22:00
← Caritade left22:00
thiago ggherdov: you'll end up with A-B-C-E-D22:00
→ RaceCondition joined22:00
cbreak ggherdov: just git fetch22:00
patagonicus ggherdov: There's git pull --rebase which will rebase your commits on the new ones upstream.22:00
linukso thiago, well22:00
cbreak ggherdov: then you'll have two heads22:00
origin/master and master22:00
linukso thiago, when you preknow the signature22:00
← RaceCondition left22:00
linukso thiago, you could use HMAC22:00
thiago linukso: "preknow" isn't a verb22:00
linukso thiago, pre-know22:00
thiago linukso: and signatures rely on encryption22:00
← angelsl left22:00
linukso thiago, not HMAC22:01
thiago, it ins't sinature itself22:01
ggherdov thiago: cbreak: thanks. patagonicus: I like this "git pull --rebase" think. trying it.22:01
thiago linukso: how do you get the HMAC?22:01
linukso thiago, with hash based methods22:01
thiago linukso: what do you compare the result to?22:01
linukso thiago, to the pre-know signature22:02
thiago linukso: how are you going to know the signature of the data that hasn't been created yet?22:02
← ErKa left22:02
jon_ if you do code reviews on a per-module basis, how do you handle changes that require multiple module changes? i'm considering a sub-module setup but this is the biggest sticking point22:02
linukso thiago, I don't, that won't be secure22:02
← HerrK left22:02
thiago linukso: so let's discard it22:02
linukso: anyway, let's get back to practical matters22:03
linukso cbreak, you know what's the problem in using quantum channels just for QKD?22:03
cbreak hmm?22:03
linukso cbreak, quantum key distribution22:03
thiago linukso: the git:// protocol allows you to transmit without encryption. It's also possible to verify that you received the data without corruption. Furthermore, it's possible to know that the data wasn't tampered with by checking the signature (you need out-of-band information for this).22:03
cbreak as I said: they are used for key transfer. They don't handle authentication, they require an out-of-band channel for coordination and they don't transfer data22:03
linukso cbreak, quantum computers, they could break the asymmetric-encryption that are made by normal computers22:03
thiago linukso: tell me: is this enough for your needs?22:04
cbreak linukso: so?22:04
← antoks left22:04
← h34d3r left22:04
thiago linukso: quantum encryption is off-topic. Please take it to another channel.22:04
linukso cbreak, is the problem in using quantum channels just for QKD22:04
← kamikaz3 left22:04
cbreak linukso: they can break about 10 bit keys22:04
linukso cbreak, they who? the computers?22:04
cbreak the quantum computers22:04
thiago linukso: take it to another channel22:04
cbreak: please don't continue this here22:05
cbreak alright.22:05
thiago linukso: are you going to answer my question?22:05
← Zarthus left22:05
→ Balliad_ joined22:05
linukso thiago, it fits, but just if I trust the channel22:05
thiago linukso: you don't have to trust the channel22:05
linukso thiago, or if the code is small enough22:05
thiago linukso: you got the data. Git by itself can verify whether it was corrupt.22:05
→ antoks joined22:06
← v0n left22:06
→ h34d3r joined22:06
linukso thiago, and what about authenticity?22:06
thiago linukso: you need out-of-band information. Like the SHA-1 that you expected to receive, or someone's public key.22:06
someone you trust22:06
jon_ if you do code reviews on a per-module basis, how do you handle changes that require multiple module changes? i'm considering a sub-module setup but this is the biggest sticking point22:06
linukso thiago, there's any GPG based git?22:06
thiago linukso: GPG can be used to sign tags and commits.22:07
cbreak jon_: just ... review multiple modules at the same time...22:07
→ blindingdawn joined22:07
← antoks left22:07
linukso thiago, yeah but automatically22:07
thiago linukso: automatically22:07
cbreak jon_: or bind them together if they make no sense alone22:07
linukso thiago, how could I do that?22:07
thiago linukso: git tag -a -s22:07
jon_ cbreak: but if i break them out into seperate repositories you could not do them at the same time, at least as far as i know22:07
→ antoks joined22:07
→ marhaban joined22:07
cbreak jon_: sure you can.22:07
submodules, subtree merging, gitslave...22:07
thiago linukso: git commit -S22:07
linukso thiago, yeah I could use git with that22:07
the protocol22:08
jon_ cbreak: right, but that would require multiple pull-requests22:08
← Balliad_ left22:08
jon_ that's where the issue is22:08
cbreak pull requests?22:08
nah22:08
← zhost left22:08
← Squarism left22:08
← Balliad left22:08
cbreak just tell them to pull the master repository22:08
that'll bring in all submodule changes too22:08
naturally, merging in a submodule is a major pain, so don't do that :D22:08
jon_ if you restrict merging into your master, evrything seems to go to hell22:09
thiago linukso: verifying a signature just takes CPU time. It might be long for big projects, but it is practical.22:09
linukso thiago, in GPG 2.1 they use ECC22:09
thiago, it'll be very much faster22:10
thiago the one that NSA was reported to tamper with22:10
jon_ i think the best i could do would be to try to approve the pull requests at near the same time, and just not worry about the brief/miniscule inconsistent state22:10
thiago in any case, that's completely irrelevant. Git uses SHA-1.22:10
linukso thiago, calm down22:10
thiago, they tamper just one implementation of it22:10
NetJunky cbreak: is it possible to checkout a specific commit as new branch. i.e. git checkout -b A SHA1?22:10
Hello71 thiago: i have reported that nsa has tampered with the earth.22:10
thiago maybe. Still, it's not relevant for this channel.22:10
cbreak NetJunky: yes. exactly that way22:10
linukso thiago, or they tampered the seed thing22:10
thiago Hello71: oh man! :-)22:11
NetJunky cbreak: okey. Just wanted to be sure first.22:11
linukso thiago, cbreak, well that's it, thank you22:12
bye22:12
← linukso left22:12
← LuxuryMode left22:12
→ FernandoBasso joined22:12
thiago ok, so all that long discussion was to find out that Git can already use GPG to sign commits and tags22:13
!next22:13
gitinfo Another satisfied customer. NEXT!22:13
cbreak things the NSA tampered with, probably (partial list): EC DRBG, the NIST curves for ECC, IPSec Standard complexity, SHA3, ...22:13
thiago resists the urge to discuss that22:13
thiago gotta go22:13
→ Droolio joined22:13
cbreak laters.22:13
→ Sam_OTnosc joined22:16
→ mizu_no_oto joined22:17
← sriracha left22:18
→ emergion joined22:18
← unreal left22:20
→ axs joined22:21
← pantsman left22:22
← milestone left22:22
→ RaceCondition joined22:23
← dreiss left22:23
← LBV_User left22:24
← Kronuz left22:24
Remram what's the standard for tag names for versions these days?22:25
0.2, v0.2, pkgname-0.2 ?22:25
→ unreal joined22:25
→ LuxuryMode joined22:25
← valgrind left22:25
← Sam_OTnosc left22:26
→ m0ltar joined22:27
→ smlgsx joined22:30
cbreak r1, r2 and so on is what I use22:31
→ Mars` joined22:31
← mFacenet left22:32
→ AWinterman joined22:33
← Sendoushi left22:34
Remram r1 ?22:35
→ sriracha joined22:35
Remram that's very unusual22:35
← amcsi left22:35
← tatsuya_o left22:36
→ Dougie187 joined22:36
→ tatsuya_o joined22:37
→ ferdev joined22:38
cbreak revision.22:39
start with 0, add one for each revision.22:39
← marhaban left22:39
← chuckharmston left22:39
→ osmosis joined22:40
← cads left22:41
Remram I could use SVN while I'm at it but that probably wouldn't help me much22:41
your scheme assumes that versions go up from one minor "revision" to another22:42
without branching, bugfixes or release candidates22:42
← tatsuya_o left22:42
Remram it is extremely limited22:42
furthermore, it doesn't really answer my question, since you could perfectly just use "1", "2", or pkgname-1 or pkgname-r122:42
→ ewnd9 joined22:44
← Norm__ left22:46
→ Risto- joined22:46
→ v0n joined22:49
→ vigs joined22:49
→ HumptyDumpty001 joined22:49
← lotus left22:50
← HumptyDumpty001 left22:50
← sriracha left22:51
cbreak Remram: branches, bugfixes or release candidates are irrelevant22:51
I just label all released versions with a release number22:51
arand__ Remram: Well, the Git documentation uses v0.2 as an example, if you're on github you might want to skip the 'v' since it adds it to the tarball name...22:51
← LuxuryMode left22:53
→ Fandekasp joined22:53
→ LuxuryMode joined22:55
← kwmiebach left22:56
→ kwmiebach joined22:57
← osmosis left22:58
← LuxuryMode left22:58
← PaulCapestany left22:58
← Amandil left22:58
→ narava_ joined22:59
→ PParth joined23:00
← ewnd9 left23:00
← PParth left23:00
← geak left23:00
← narava_ left23:01
→ narava_ joined23:01
→ PParth joined23:01
← stamina left23:02
← staafl left23:04
← j416 left23:04
→ tatsuya_o joined23:07
← _Harm_B_ left23:07
→ _Harm_B_ joined23:08
→ LuxuryMode joined23:09
← nwertman_ left23:09
→ cmpera joined23:11
→ j416 joined23:14
← tatsuya_o left23:16
← narava_ left23:17
← v0n left23:18
→ narava joined23:19
← petey left23:19
← AWinterman left23:20
→ AWinterman joined23:20
← AWinterman left23:21
→ AWinterman joined23:21
→ Gamecubic_ joined23:21
→ matthiaskrgr joined23:23
gitinfo set mode: +v23:23
← narava left23:23
matthiaskrgr >Be careful not to step in the git-gui - nobody gc'ed after the party last night23:23
what?23:23
offby1 broken glass everywhere people xxxxxing on the stage you know they just don't care23:24
← Gamecubic left23:24
matthiaskrgr mh23:24
offby1 It's like a jungle. Sometimes it makes me wonder how I keep from going under.23:24
matthiaskrgr is there a way to limit git gc on memory (ram) it uses?23:24
offby1 apart from ulimit, I doubt it23:25
→ vad joined23:25
vad How could I make a bare repository keep a reflog?23:25
grawity man git config – look for the pack.* settings23:25
gitinfo the git-config manpage is available at http://jk.gs/git-config.html23:25
matthiaskrgr I have a repo of around 5 GB and only 3.5 GB ram23:25
← epicowl left23:25
grawity vad: core.logAllRefUpdates23:25
vad cool.23:26
offby1 I'd have thought a bare repository _would_ keep a reflog23:26
← oriba left23:26
vad matthiaskrgr: git gc --max-pack-size=500m?23:26
offby1: not by default.23:26
matthiaskrgr error: unknown option `max-pack-size=500m'23:27
vad it's for git repack23:28
`git gc` is `git repack -adf` or thereabout, IIRC23:28
matthiaskrgr i tried git repack but it seemed to make the pack size even bigger O.O23:28
grawity a plain `git repack` doesn't delete loose or duplicate objects by default23:28
so, did you look through the git-config manpage for pack.* settings?23:29
← axs left23:29
→ dreiss joined23:29
vad follow `git repack` by `git prune` then :)23:29
← theRealGentleman left23:31
← buhman left23:32
← rurufufuss left23:33
matthiaskrgr yes, reading it23:33
← FernandoBasso left23:34
→ LucasTizma joined23:35
← KillYourTV left23:36
← dreiss left23:37
← tagava left23:37
→ lotus joined23:38
← falconfox left23:41
→ chuckharmston joined23:42
→ KillYourTV joined23:42
← King_Hual left23:43
← NetJunky left23:43
→ notjohn joined23:43
→ theRealGentleman joined23:47
← chuckharmston left23:47
→ sriracha joined23:48
← mixel left23:50
← m0ltar left23:51
→ psyke83 joined23:51
matthiaskrgr can I put comments into my .gitconfig file?23:52
psyke83 hey guys. I'm updating a diverged branch using "git rebase --onto temp oldrevision latestrevision"; my problem is when the first merge failure occurs, "git mergetool" shows that no files need merging. How can I fix this?23:52
matthiaskrgr ok found something :)23:53
← blindingdawn left23:53
→ blindingdawn joined23:53
← AWinterman left23:54
→ AWinterman joined23:55
← blindingdawn left23:58
← LuxuryMode left23:58
← lotus left23:59

Logs Search ←Prev date Next date→ Channels Documentation