IRCloggy #git 2013-06-09

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.

2013-06-09

← laissez-f left00:00
← mmlac left00:00
← conan left00:00
→ ktosiek joined00:00
← knite left00:01
→ Balliad joined00:03
← fridim_ left00:03
← pierre1 left00:05
← infogulch left00:06
← DonRichie left00:06
→ DonRichie joined00:07
← nicedice left00:07
→ m0ltar joined00:08
→ infogulch joined00:09
alfplayer Hello. What can I use to keep a couple of images out of a repo? I see many possible solutions and I don't know where to start.00:09
← skfax left00:10
→ wrouesnel joined00:11
← rangle left00:11
← chris2 left00:12
→ rickruby joined00:14
offby1 alfplayer: how about not adding them in the first place?00:14
→ snearch joined00:14
→ chris2 joined00:14
alfplayer offby1: Yes, but some tracking may be good.00:15
← devoldmx left00:15
→ devoldmx joined00:15
← wrouesnel left00:15
alfplayer It can be useful to tell contributors when they are updated.00:15
→ wrouesnel joined00:16
offby1 well, as far as I know, "tracking" requires that they be present in the repository. It's pretty much all or nothing.00:16
unless ...00:16
you use something like ... uh ... Joey Hess' thing for storing ginormous files in git00:16
whose name escapes me00:16
alfplayer You mean git-annex, but it seems overkill.00:16
offby1 ya00:17
only thing I can think of.00:17
alfplayer I see things like git-media and git-fat.00:18
Or any kind of subproject.00:18
But how do I choose?00:18
← rickruby left00:18
← kfogel left00:19
offby1 no idea.00:19
← Noldorin left00:20
→ GarethAdams joined00:20
← asteros left00:21
→ SuperrMann joined00:23
→ rangle joined00:23
orion_ .00:23
← orion_ left00:23
→ JaredMiller joined00:24
← rangle left00:24
→ mmlac joined00:24
← arand left00:28
← etcetera left00:29
← Balliad left00:30
← momomomomo_ left00:30
← wrouesnel left00:31
← BJfreeman left00:31
← kenperkins left00:32
→ _BJfreeman joined00:32
→ tmcmahon joined00:33
_BJfreeman → BJfreeman00:33
bremner !media00:34
gitinfo [!annex] git-annex and git-media are two solutions to the !binary problem. They work by keeping the blobs outside of the repo, storing a reference to the blob in the repo instead. See http://git-annex.branchable.com/ and https://github.com/schacon/git-media00:34
bremner !binary00:34
gitinfo Storing binary files in git causes repo balloon, because they do not compress/diff well. In other words, each time you change a file the repo will grow by the size of the file. See !annex for some solutions00:34
← Dougie187 left00:34
alfplayer Not any subproject solution?00:34
→ kenperkins joined00:35
bremner what do you mean?00:35
alfplayer I'd like to avoid cloning old binaries.00:36
E.g. I see someone in the web is suggesting submodules.00:36
← chamunks left00:36
bremner how would that help?00:36
CareBear\ !blog00:37
gitinfo Blog posts, while helpful and informative, are quite often outdated, give bad advice, or are just plain wrong. Please don't rely solely upon them, or treat them as authoritative.00:37
alfplayer To optionally avoid checking out the binaries?00:37
The recommendation about submodules is from http://programmers.stackexchange.com/a/8129900:38
cmn that won't stop you from having to download older binaries00:39
if you don't want to check them into version control, then you'll download them; if you don't, git-annex will give you some level of tracking while not making you download them00:39
← dougsland left00:39
← GarethAdams left00:40
cmn erm, s/if you don't/if you do/00:40
alfplayer That would force contributors to use git annex? Maybe it can be avoided?00:41
→ tjbiddle_ joined00:41
cmn by checking in files into git00:41
→ chamunks joined00:41
alfplayer It's not in my distro's repos.00:42
Is it worthy to look at git-media?00:42
← tjbiddle left00:43
tjbiddle_ → tjbiddle00:43
cmn not according to you00:43
alfplayer Oh, ok.00:43
It's not in the repos either, if that's what you mean.00:43
← jcao219_ left00:45
→ tjbiddle_ joined00:46
← tjbiddle left00:47
tjbiddle_ → tjbiddle00:47
→ CodeShark joined00:47
← tg left00:50
← JaredMiller left00:51
← snearch left00:51
← pastime left00:52
→ Nicholas_Rishel joined00:53
→ DOCRukuss joined00:53
← infogulch left00:54
→ bartek joined00:54
← SuperrMann left00:56
→ R1KRDO joined00:56
← R1KRDO left00:57
← Nicholas_Rishel left00:59
← martijnHH left00:59
→ vitor__ joined01:00
← D9 left01:01
→ tg joined01:03
← tg left01:03
→ tg joined01:03
→ ellio joined01:03
ellio Is there a way to add an entry to the reflog by hand, or something like that?01:03
← vitor_ left01:04
ellio I'm still a complete reflog noob. But I just made some rather hardcore, dangerous, sweeping changes to my repository … and right now, the *old* heads (before all the changes) of the three branches that got changed thoroughly, are still on my remote. But I'm about to force-push over that, as well, and I'm kinda hoping for a way to record those old three01:04
heads for posterity in my reflog, incase I realize a week down the road that I'd screwed this up subtly.01:04
→ kcj joined01:05
alfplayer They are a very small amount of binaries for now, so I think I'll try to avoid tracking them in the project's repo for now.01:05
→ lamarus joined01:05
alfplayer Thanks for the help.01:05
→ fridim_ joined01:06
→ pastime joined01:07
→ Vampire0_ joined01:07
← PaulCapestany left01:09
← spaceonline left01:09
← Vampire0 left01:10
→ douglaslassance joined01:11
→ dougsland joined01:14
← ntus1017 left01:16
→ asteros joined01:17
← tadasZ left01:17
offby1 ellio: I've never heard of hacking a reflog by hand; that's sorta against the idea of the reflog.01:19
Its purpose is to record the values that various refs have taken on at various times.01:19
→ huoxito joined01:19
offby1 if you're worried about certain commits becoming unreachable, just tag 'em01:20
← pastime left01:20
→ drewery joined01:21
drewery Hello Everyone01:21
I have a question.01:21
→ PaulCapestany joined01:21
drewery How can I get git ignore a folder and it is subfolders.01:21
→ JaredMiller joined01:21
drewery ?01:21
There is one upload folder that also creates subfolders by date. I want to get GIT ignore the folder and any addtional subfolder may be created in the future.01:22
Is there a way to do that?01:22
Thanks in advance..01:23
← tmcmahon left01:23
offby1 drewery: I think you just put the folder's name into .gitignore01:25
Mamoru drewery: add folder/* to .gitignore01:25
drewery ok my gitignore has * and !.gitignore lines01:25
Mamoru Just add a new line01:26
drewery so I would only add folder/*01:26
ok thanks Mamoru !01:26
and thanks offby1 as well01:26
Mamoru yes, folder being the name of the folder with the subfolders you want to hide ;)01:26
You're welcome01:26
drewery oh Mamoru here is the thing01:26
← BJfreeman left01:27
Mamoru the * will make sure all subfolders are automatically ignored01:27
drewery this is an upload folder and it creates new subfolders with different names every month01:27
Mamoru ^01:27
→ LuxuryMode joined01:27
drewery is there a way to tell git to ignore all the folders with anyname should be ignored01:27
Mamoru */anyname/*01:28
drewery :)01:28
ok I will try */*01:28
kpreid erm,01:28
Mamoru */*will ignore everything!01:28
kpreid if you have one folder 'folder' in the root of your repos then all you need to do is put '/folder' in the gitignore01:28
that will ignore all its contents01:29
(leading / because gitignore is recursive by default)01:29
Mamoru oh, right, the * is optional01:29
→ intripoon joined01:29
Mamoru you can add cache/* if you do want the cache folder to be in the repo but not its contents01:30
kpreid no, git does not track folders themselves. that makes no difference01:30
Mamoru no?01:31
drewery awesome... double checking : if the folder is under 'folder1/folder2/folder3/' and the location of the .gitignore being under folder3 I would type 'folder1/folder2/folder3/' and leave it like that01:31
correct?01:31
→ arand joined01:31
kpreid drewery: no, /folder1/folder2/folder3/01:31
drewery right! thanks for the correction01:32
Mamoru Hmm.. I'm sorry.. I seem to be confused. My mistake. Please listen to what kpreid says instead ;)01:32
← breck7 left01:32
← intripoon_ left01:33
drewery Mamoru: I think you also had a logical approach thinking that only files are being tracked not folders, so thank you for your effort as well01:33
and thanks kpreid01:33
→ bitkiller_ joined01:34
Mamoru Actually, I was mixing up different version control systems.. There are way too many out there :|01:35
→ jalmeida joined01:35
← JaredMiller left01:36
douglaslassance → Asterokid01:37
← jalmeida left01:37
→ BiggFREE joined01:37
→ jalmeida joined01:37
bremner there is only one.01:37
→ pierre1 joined01:37
bremner the rest are illusions.01:37
← asteros left01:37
← lifelike left01:38
→ asteros joined01:38
→ nbouscal joined01:40
← asteros left01:41
→ erkules joined01:41
← bitkiller left01:42
bitkiller_ → bitkiller01:42
Mamoru Currently, I (have to) use cvs, svn, bzr, hg and git01:43
drewery yeah, if you have different projects being thrown at you01:43
← Cykey left01:43
← erkules_ left01:44
drewery I got this project with SVN and I had no idea. GIT has been the only thing I showed an effort to learn and still learning and extremely happy about it. Since I didn't use any other I have no judgement of comparison however.01:44
Mamoru I'm not even bothering to mention Dropbox and M$ Team Foundation Server as 'version control systems'01:45
drewery oh, I use dropbox but for simple tasks and didn't know that it has version control stuff with it01:46
→ tjbiddle_ joined01:46
Mamoru up to 15 or 30 days iirc01:46
on a free account01:46
← Merwin_ left01:46
Mamoru mergin is a nightmare.. xD01:46
drewery by the way do I have to stage the .gitignore file for it to be effective?01:46
→ breck7 joined01:47
→ Merwin_ joined01:47
drewery I used absolute path, server root path, git initiated root path but this doesn't seem to work for me01:47
Mamoru It's my illusion that you should01:47
← m0ltar left01:47
drewery I also used touch to create a test.file file01:47
just to see if the file will disappear from git status list01:48
I am getting no results so far01:48
I must be doing something wrong01:48
← tjbiddle left01:48
← LuxuryMode left01:50
drewery ok ok ok problem solved01:51
I deleted everything and just typed *01:51
and all the subfolders together with the test file I created disappeared from the untracked files list01:52
← tjbiddle_ left01:52
drewery that was easy! (not when you don't know it ! )01:52
Mamoru well.. * ignores everything01:52
← nbouscal left01:53
→ trusktr joined01:53
drewery Mamoru: My mistake was I was assuming it only ignores files even I use unix and knew the command... oh well, I think I am working too much lately01:54
→ tjbiddle joined01:54
Mamoru hmm01:55
Now I'm not so sure I should even answer questions anymore :p01:55
Its almst 4 am01:55
Maybe I should sleep :p01:55
→ dhruvasagar joined01:56
→ noxd joined01:56
← Merwin_ left01:56
→ Merwin_ joined01:57
drewery You should be somewhere between england and greece :P01:59
→ LuxuryMode joined01:59
→ tjbiddle_ joined02:01
← tallest_red left02:01
drewery Also just found out that if I am working in a folder that is ignored and I call git to stage changes (that even has nothing to do with ignore folder) it refuses to stage stuff. I get out of the ignored folder and it works just fine... interesting..02:01
← tjbiddle left02:01
tjbiddle_ → tjbiddle02:01
→ chrisshattuck joined02:04
← chrisshattuck left02:04
→ tallest_red joined02:05
→ asteros joined02:05
← Markvilla left02:08
→ luc_ joined02:09
← vitor__ left02:12
→ hello798 joined02:12
→ vitor__ joined02:13
→ angelsl joined02:13
→ Cykey joined02:16
← CodeShark left02:16
→ CodeShark joined02:18
→ jaequery joined02:24
← DonRichie left02:25
← breck7 left02:25
← devoldmx left02:27
← dhruvasagar left02:27
ELLIOTTCABLE I have a rather unique problem:02:32
I need to copy the *metadata* from one commit, to one I'm about to make.02:32
Or, more accurately: I need to split somebody else's commit in two, in the middle of an interactive-rebase ... but I need the second commit *as well as the first* to retain his authorship information (most notably date, although his name should be retained as well)02:32
← LuxuryMode left02:32
ELLIOTTCABLE basically: in the history, it should look like he made two commits at the time he made the original commit, after I'm done with my splitting02:33
how do? :D02:33
→ LuxuryMode joined02:34
← vitor__ left02:34
→ vitor__ joined02:35
→ devoldmx joined02:36
→ leeN joined02:36
→ dhruvasagar joined02:36
→ Jahm joined02:37
← setmeaway2 left02:37
→ btree joined02:37
offby1 first, unstick your caps lock key02:38
anyway, git-commit has a pile of options for snagging metadata from one commit to the new one02:38
-C gets a log message, e.g.02:39
← btree left02:39
offby1 you can set --author and --date "by hand"02:39
Not sure there's anything else02:39
← plumpkin left02:39
← aard left02:39
← chamunks left02:39
→ tubbo joined02:40
→ chamunks joined02:40
← Dave^| left02:42
← lamarus left02:43
← noxd left02:44
→ mechanicalduck joined02:47
← phipes left02:48
ELLIOTTCABLE hm, has to be by hand02:49
was afraid of that. Ended up already doing that, anyway, but feels like there'd be a better way to do that.02:49
← dhruvasagar left02:50
← chamunks left02:56
→ chamunks joined02:56
← Davey left02:59
← lacrymology left03:00
← leeN left03:01
offby1 I can't think of any fundamentally different way.03:01
tubbo hey guys03:02
when i push to my git remote, am i using `git-shell` interactively?03:02
(like, is that how git is transferring files between my machine and the remote host?)03:03
→ phipes joined03:03
← Alz_ left03:04
← donmichelangelo left03:04
→ donmichelangelo joined03:05
→ mrballcb joined03:05
→ dhruvasagar joined03:05
→ preflex_ joined03:06
← duthils left03:07
← preflex left03:07
preflex_ → preflex03:07
← chamunks left03:09
→ kfogel joined03:09
← drewery left03:10
→ chamunks joined03:10
→ leeN joined03:11
→ Davey joined03:12
← TomyLobo left03:13
offby1 no, you aren't03:14
not even sure what git-shell is, actually.03:14
hm, maybe you are, indirectly.03:14
I've been using git for years and never think about git-shell (as you can see I wasn't even aware of what it was)03:15
that doesn't mean you're not using it though.03:15
← fluxdude left03:17
← freestyl3r left03:17
← dhruvasagar left03:17
← hellopat_ left03:19
← devoldmx left03:19
→ mribeirodantas joined03:19
→ simesy joined03:20
→ devoldmx joined03:20
→ Runeg joined03:20
← troyt left03:22
← LuxuryMode left03:23
← jaequery left03:24
← simesy left03:24
← Cykey left03:26
tubbo yeah me neither03:26
there also doesn't seem to be many docs related to it03:26
offby1 just outta curiosity: why do you ask?03:27
← thiago left03:28
← mmlac left03:28
→ weiyang joined03:28
tubbo offby1: i'm trying to build something that sits between git-shell and the client, that will authenticate users based on the given SSH private key to my database of users03:29
sorta like gitolite except without requiring me to update its list of known keys03:29
instead, i'll make a database call (well, an HTTP request..)03:29
offby1 yow03:29
→ simesy joined03:29
← simesy left03:29
brainproxy_ → brainproxy03:30
offby1 tubbo: if I were you, I'd just set up a "git server" (it could be the same machine you're already using) and start poking at it.03:30
Presumably one of the access methods will use it ... probably ssh03:30
tubbo: I'd also ask sitaram for advice. He'll either 1) tell you that gitolite can already do what you want; 2) offer to tweak it for you; or 3) give you advice on building your own thang03:31
tubbo offby1: what do you mean by "poking at it"?03:32
and yeah, i do probably need to talk with sitaram :)03:33
→ mencito joined03:33
tubbo i'm a big fan of gitolite and would definitely rather use a tried and tested tool rather than something i have to build03:33
→ zheng joined03:34
→ hwrd|work joined03:34
offby1 tubbo: I just meant: push to it, and observe. See if you can tell if it's starting a git-shell process.03:35
tubbo oh ok03:35
offby1 Read the docs and see if you need to do something to enable git-shell (I suspect you do)03:35
→ rchavik joined03:35
tubbo oh it's `git daemon`03:35
not `git server` :P03:35
now i understand03:36
→ breck7 joined03:36
tubbo offby1: well, i'm not really concerned with controlling public git:// access to my repos, i'm more concerned with SSH access03:37
isn't that what `git daemon` does? provides git protocol access to a given repo?03:37
offby1 yep, read-only if I recall correctly.03:37
→ jaequery joined03:38
tubbo right, so i think that part is covered.03:38
well actually, i'll probably have to see what's up with that03:38
→ simesy joined03:38
sitaram tubbo: the problem you will have is this piece (quoting from what you said): "based on the given SSH private key"03:39
sshd does not give you that information03:39
tubbo ah crap03:39
:)03:39
i wonder how i can figure out which user is trying to access my server..03:39
← leeN left03:40
← rchavik left03:40
→ thiago joined03:40
sitaram tubbo: if it's via ssh, there really *is* no way other than what gitolite does, (or patch sshd)03:40
← breck7 left03:40
tubbo sitaram: and with gitolite, the only way i can actually make changes is via that gitolite-admin repo, right?03:41
sitaram well... normally yes. But there are other ways03:41
you can ignore the whole admin repo and directly manipulate ~/.gitolite/conf and ~/.gitolite/keydir03:42
then run 'gitolite compile; gitolite trigger POST_COMPILE'03:42
tubbo oh cool03:42
so like, when a user makes a new repo i could regenerate ~/.gitolite/conf, and when a new user is added i can regenerate ~/.gitolite/keydir03:43
→ echo-area joined03:43
tubbo then i can use gitolite to manage access for me03:43
which would be super cool03:43
← asteros left03:43
sitaram well, ~/.gitolite/conf/gitolite.conf is the file but the idea is correct03:44
tubbo oh right03:44
and i'm running gitolite compile.. on the machine i've installed gitolite on, right?03:44
or wait, there is no client machine :)03:44
→ DolphinDream joined03:44
sitaram back to ssh; see https://groups.google.com/d/msg/gitolite/MX4qQ0LqWNs/QmvxS_okWLMJ for a quick explanation (from me) of how one of openssh's new features *could* have been better for (y)our purposes03:44
tubbo: all 'gitolite <subcommand> <args>' is on the gitolite server, as the "git" (or maybe "gitolite") user03:45
i.e., whatever user is the "hosting" user03:45
tubbo oh ok03:46
sitaram wonders... what it would take to pass the fingerprint as the second argument (even if the user name is the first one)03:47
tubbo that would be pretty cool but i understand like, why they wouldn't do that03:47
sitaram why?03:47
tubbo then again, when was the last time you gave out your private key to a host you didn't trust :)03:47
sitaram huh?03:47
key *fingerprint*03:47
not private key03:47
tubbo yeah03:47
i guess i was just paranoid there for a second ;)03:48
that would definitely make this whole process a lot easier03:48
← hwrd|work left03:49
tubbo in the future we might have to hack sshd, but i really don't want to.03:49
sitaram user connects to git@host. Offers a key. Ssh 6.2 passes "git" to the command named in this new option. The program uses that to look up some master database of pubkeys and passes all relevant ones in the format of an ssh authkeys file.03:49
all I'm saying is that at that time sshd pass the FP as arg-2 so the search can be narrowed!03:50
← penk left03:50
→ NiGHTMAN joined03:51
← simesy left03:51
tubbo hmm, that's another strategy03:52
i could just have AuthorizedKeysCommand generate a key file, and have it be empty if i don't want the user to get access to the repo03:52
that's as long as SSHd is passing the path of what the client wants to access03:53
← NiGHTMAN left03:53
tubbo that way i could look it up by "user/repo.git", which is how we have things organized.03:53
→ R1KRDO joined03:54
tubbo however, i still don't know the key03:55
oh wait i'm inventing gitolite LOL03:55
reinventing*03:55
→ rchavik joined03:56
offby1 suspected as much03:57
sitaram tubbo: I've just subscribed to the openssh user mailing list and am going to ask if the FP can be added somehow. Not even sure if the info is available at the time that the authorizedkeyscommand helper is called though03:57
← HelloComputer15 left03:58
offby1 I wonder if you could coax that outta $SSH_AUTH_SOCK somehow03:59
← donmichelangelo left04:00
→ dhruvasagar joined04:00
→ donmichelangelo joined04:00
← zheng left04:04
→ zheng joined04:04
→ _BJfreeman joined04:05
→ asteros joined04:06
_BJfreeman → BJfreeman04:06
→ curvedmark joined04:08
← mrballcb left04:09
← devoldmx left04:09
← jpmelos left04:10
→ devoldmx joined04:10
← solocommand left04:10
→ solo-work joined04:10
→ sazzq joined04:11
→ Dreamer3 joined04:14
← devoldmx left04:16
→ devoldmx joined04:16
← diegoviola left04:18
← dhruvasagar left04:20
→ penk joined04:21
→ pabs3 joined04:24
pabs3 is there any way to make git tag always sign tags when creating them?04:24
← jaequery left04:26
← n8chur left04:26
→ n8chur joined04:28
→ r0bby joined04:28
→ jaequery joined04:28
← penk left04:29
→ ezrafree joined04:29
→ iodarkstar joined04:31
→ Tommy[D]_ joined04:34
← Joshi3s` left04:34
← ag4ve left04:35
→ ag4ve joined04:35
← asteros left04:36
← phipes left04:36
← DolphinDream left04:38
← Tommy[D] left04:38
← PaulCapestany left04:39
→ jpmelos joined04:45
→ chrisshattuck joined04:49
← chrisshattuck left04:49
→ dsdeiz joined04:50
← the_cybersphinx left04:50
→ PaulCapestany joined04:50
← kukks left04:52
→ yuwang joined04:53
← iodarkstar left04:54
offby1 pabs3: probably the easiest way is to make an alias, so that you don't have to type "-s"04:54
pabs3 so just tag = tag -s in the [alias] section of ~/.gitconfig?04:56
← R1KRDO left04:56
→ simesy joined04:56
← chamunks left04:57
→ chamunks joined04:58
← moondoggy left05:02
SamB pabs3: I heard some silly person banned aliasing existing commands :-(05:03
also what's with the 305:04
→ private joined05:05
← sazzq left05:06
pabs3 SamB: confirmed, the alias doesn't work. pabs3 cause someone else was pabs on freenode before me05:08
→ Sigma joined05:09
→ moondoggy joined05:11
← moondoggy left05:11
→ moondoggy joined05:11
← pierre1 left05:12
sitaram the restrictions against aliasing existing commands is because scripts can go haywire if some options get turned on behind the scenes05:13
SamB hmm05:13
→ Dead_Body joined05:13
gitinfo set mode: +v05:14
SamB presumably the lack of a crystal-clear boundary between plumbing and porcelein doesn't help here :-(05:14
← chamunks left05:15
→ jcao219 joined05:16
← weiyang left05:16
← simesy left05:19
← mckendricks left05:19
offby1 I'd just as soon my plumbing remain opaque, thanks very much.05:20
→ vedant joined05:21
← madewokherd left05:22
SamB guh05:23
→ simesy joined05:24
→ irqq joined05:27
← simesy left05:27
← lonewulf` left05:28
→ iodarkstar joined05:28
→ d2dchat joined05:31
← iodarkstar left05:31
← SBoudrias left05:31
→ arietis joined05:31
← avinashrbhat left05:32
← dsantiago left05:33
→ dhruvasagar joined05:35
← vedant left05:38
→ simesy joined05:38
← simesy left05:39
→ avinashrbhat joined05:43
→ lindenle joined05:46
← BJfreeman left05:49
← arietis left05:49
→ jceb joined05:50
→ homer91 joined05:50
→ hide_ joined05:50
← hide left05:51
SeoZ-work[AWAY] → SeoZ05:53
← galaris left05:53
← takezawa left05:55
→ iodarkstar joined05:56
← barneybook left05:56
EugeneKay SamB - https://madeitwor.se/scripts/blob/master/bash/git-wrapper.sh ;-)05:59
← huoxito left05:59
→ barneybook joined05:59
→ guiambros joined06:03
← m0viefreak left06:03
→ Alz_ joined06:04
→ mmlac joined06:05
← tanath left06:05
→ longbeach joined06:07
← Sigma left06:09
→ martijnHH joined06:11
← kfogel left06:12
← mmlac left06:12
← iodarkstar left06:12
← rchavik left06:14
← lindenle left06:14
← FreezingCold left06:16
→ hoschi joined06:18
→ mechanicalduck_ joined06:19
← pygmael left06:19
→ rchavik joined06:19
← mechanicalduck left06:22
← fridim_ left06:25
→ skystar joined06:26
→ skystar84 joined06:26
← skystar84 left06:27
← skystar left06:27
→ PaulePanter joined06:28
→ Cromulent joined06:28
← Asterokid left06:31
→ wrouesnel joined06:31
← martijnHH left06:33
→ FreezingCold joined06:33
← PaulCapestany left06:33
→ PaulCapestany joined06:40
→ douglaslassance joined06:45
→ toobluesc joined06:45
→ ej joined06:45
← mribeirodantas left06:46
→ wwu joined06:47
→ Jackneill joined06:49
← Jackneill left06:49
→ Jackneill joined06:49
← Jahm left06:49
← douglaslassance left06:50
→ charliesome joined06:51
→ mechanicalduck joined06:52
← wwu left06:52
← mencito left06:52
→ Jahm joined06:53
→ wwu_ joined06:54
← mechanicalduck_ left06:54
← wrouesnel left06:54
← dhruvasagar left06:56
→ dhruvasagar joined06:57
← hello798 left06:59
← jcao219 left07:00
← private left07:01
→ elixir joined07:03
← OOPMan left07:05
→ douglaslassance joined07:08
← jceb left07:08
← relipse left07:08
→ darnir joined07:09
← darnir left07:09
→ darnir joined07:09
← konundra left07:10
← guiambros left07:11
← rasto left07:12
→ chamunks joined07:13
← zheng left07:20
← ToothlessGear_ left07:21
→ ToothlessGear_ joined07:21
ToothlessGear_ → ToothlessGear07:23
→ Blahhh joined07:24
→ rasto joined07:26
← elixir left07:26
→ r0bby_ joined07:30
← baneM left07:31
→ nerdfiles joined07:31
→ baneM joined07:31
→ legumbre_ joined07:32
→ samir joined07:33
→ zizzfizzix_ joined07:33
← zizzfizzix_ left07:33
→ zizzfizzix_ joined07:33
→ Merwin__ joined07:33
→ deavidsedice joined07:33
→ jayne| joined07:33
→ mlb-__ joined07:35
→ catsup joined07:35
→ harv_ joined07:35
→ lb joined07:35
→ s0x_ joined07:35
→ padhah joined07:35
→ kortti_ joined07:35
→ JakePee_ joined07:35
→ cxreg2 joined07:35
→ tonyj_ joined07:35
→ unreal_ joined07:35
→ banzoune1 joined07:35
← nigelb left07:35
→ iveqy joined07:35
→ vinc_ joined07:35
→ chuckf_ joined07:35
→ nigelb joined07:35
→ shabble_ joined07:35
← shabble_ left07:35
→ shabble_ joined07:35
→ jabular_ joined07:35
→ charliesome_ joined07:36
→ DrNick2 joined07:37
→ Jahm_ joined07:37
→ iron_hou1 joined07:37
→ ecube joined07:37
→ lrvick3 joined07:37
→ ping_pong joined07:37
← avinashrbhat left07:37
→ satiricon_ joined07:37
← FreezingCold left07:37
→ Vampire0 joined07:37
→ specing_ joined07:38
→ farn_ joined07:38
← charliesome left07:38
← Mpenzi left07:38
charliesome_ → charliesome07:38
← lb_ left07:38
← coopsh left07:38
← ThomasCompix left07:38
← mlb-_ left07:38
← sniffell1 left07:38
← lyang0 left07:38
← rgrinberg left07:38
← padhah_ left07:38
← iron_houzi left07:38
← beawesomeinstead left07:38
← arnihr left07:38
← iveqy_ left07:38
← Guest96500 left07:38
← nerdfiles1 left07:38
← specing left07:38
← Guest975 left07:38
← jday_ left07:38
← ping-pong left07:38
← alexander__b left07:38
← kortti left07:38
← tonyj left07:38
← cxreg left07:38
← Nevik left07:38
← jjore left07:38
← satiricon left07:38
← vinc left07:38
← eli left07:38
← JakePee left07:38
← csgeek left07:38
← lrvick1 left07:38
← harv left07:38
harv_ → harv07:38
← s0x left07:38
→ rgrinberg1 joined07:38
← chuckf left07:38
← banzounet left07:38
← Jahm left07:38
← yuwang left07:38
← r0bby left07:38
← Merwin_ left07:38
← Vampire0_ left07:38
← jabular left07:38
← zizzfizzix left07:38
← jfinn left07:38
← farn left07:38
← DrNick left07:38
← unreal left07:38
Jahm_ → Jahm07:38
← t_hash left07:38
← deavid left07:38
← shabble left07:38
← legumbre left07:38
← jayne left07:38
→ Mpenzi joined07:38
→ jjore joined07:38
← Cromulent left07:38
→ jfinn joined07:38
→ Nevik joined07:38
→ yuwang joined07:38
→ hash joined07:38
→ sniffell1 joined07:38
specing_ → specing07:38
→ alexander__b joined07:39
→ shruggar joined07:39
← chamunks left07:40
→ rihnapstor joined07:40
→ coopsh joined07:40
→ lyang0 joined07:40
→ chamunks joined07:41
→ jayfk joined07:43
← tg left07:43
→ abstrusenick joined07:43
← Goplat left07:43
→ pantsman joined07:45
← pantsman left07:45
→ pantsman joined07:45
→ avinashrbhat joined07:46
→ dsantiago joined07:46
→ stamina joined07:47
← shruggar left07:47
→ babilen joined07:48
→ daanavitch joined07:50
→ tg joined07:51
← echo-area left07:51
← dsdeiz left07:53
→ rudak1 joined07:55
tjbiddle Anyone available this late? I have a colleague who can clone in a project via `git clone` but when that same project is reference as a submodule and he goes to `git submodule init && git submodule update` he is prompted for a password07:56
I'm assuming `git submodule` for some reason is checking a different path07:56
He's on a Windows box.07:56
different path for the SSH key *07:56
→ echo-area joined07:57
jayne| → jayne07:58
← causasui left07:58
cbreak tjbiddle: it uses the same backend as the normal repository07:58
the same backend as every repository07:59
→ causasui joined07:59
tjbiddle cbreak: By 'backend' do you mean `git submodule` should essentially just be calling out to `git clone` ?07:59
cbreak the problem must be that his ssh key is either not found, or not accepted by the remote07:59
← pabs3 left07:59
tjbiddle Minus some special git submodule stuff07:59
cbreak tjbiddle: it will,07:59
(git fetch)08:00
tjbiddle cbreak: Yeah - that's what I was assuming. But I don't know why one accepts his ssh key perfectly fine, and the other doesn't08:00
cbreak check the remote config08:00
look at .git/config in both the submodule and the outer repository08:00
and look at .gitmodules08:00
← chamunks left08:03
→ chamunks joined08:04
← tplaner left08:05
zz_pinage404 → pinage40408:06
pinage404 → pinage404afk08:06
→ DrBenway joined08:08
→ aartist joined08:08
DrBenway hi, so i have a master branch on github which i've clone and have been working on locally. i then created a branch1 and i would like to push my local commits to it instead of master08:09
how can i do this08:09
git push origin branch1 doesn't seem to work08:09
→ kerrick_ joined08:09
DrBenway normally i push with git push origin master08:09
← thiago left08:09
→ rnz_ joined08:10
pinage404afk → zz_pinage404afk08:10
→ zz_pinage404 joined08:10
← rnz_ left08:10
cbreak DrBenway: git push -u origin branch1 should work08:11
if it doesn't pastebin the whole command and result08:11
← tjbiddle left08:11
→ elixir joined08:12
→ paul424 joined08:13
DrBenway it gives me the same as git push origin branch108:13
result:08:13
error: src refspec opengl3 does not match any.08:13
error: failed to push some refs to 'https://github.com/fcoulombe/kinevox.git'08:13
(opengl3 being my branch)08:13
it's like if it's not picking up that the branch exists on origin08:15
(you should be able to see by the link that it does indeed exist)08:15
unless i've messed up its creation08:15
→ wrouesnel joined08:15
cbreak DrBenway: it doesn't matter wether it exists on origin08:15
what matters is that it exists as local branch08:16
pastebin git branch -a08:16
DrBenway: if it is not listed as local branch, then you failed to create the branch08:16
← douglaslassance left08:16
DrBenway it is in the list08:17
remotes/origin/opengl308:17
cbreak show it.08:17
that's not a local branch...08:17
DrBenway it's not the one active but it's there08:17
cbreak that's a remote tracking branch08:17
→ hinst joined08:17
cbreak those are irrelevant for this problem08:17
so... conclusion: you did NOT create the branch08:17
create it08:17
THEN you will be able to push08:17
DrBenway erm... how?08:17
cbreak branches are created with git branch or git checkout -b08:18
(I prefer the latter)08:18
DrBenway: the branch with the * in front is your current branch08:18
→ jceb joined08:18
→ Cromulent joined08:19
cbreak git checkout -b newbranchname will switch it to the new branch you create08:19
DrBenway alright got it thanks08:19
farn_ → farn08:20
cbreak after that, use git push -u origin branchname once08:20
and from then on only git push while on that branch08:20
(with git config push.default upstream)08:20
→ Sigma joined08:22
→ maslowbeer joined08:22
← rihnapstor left08:22
iveqy I've followed the discussions on the git mailing list. And it seems like git is sooner or later going to be replaced by libgit2 internally. Is it even worth to continue to write on git.git if it's probably going to be replaced anyway?08:22
← rchavik left08:23
cbreak switching to sha3 would be nice.08:24
→ gittish joined08:25
→ madprops joined08:27
→ tjbiddle joined08:28
← dhruvasagar left08:29
← tg left08:32
← tjbiddle left08:32
→ rihnapstor joined08:35
← rihnapstor left08:35
→ OOPMan joined08:36
→ gasbakid joined08:36
→ thierryp joined08:37
← skorgon left08:40
← kerrick_ left08:40
← akiress left08:41
→ tg joined08:42
← ELLIOTTCABLE left08:42
→ ELLIOTTCABLE joined08:42
→ howlymowly joined08:42
← thierryp left08:43
→ takezawa joined08:43
← Cromulent left08:45
← boombatower left08:47
→ vedant joined08:48
← archdori1 left08:49
← w0ng left08:51
→ _iron joined08:51
→ dhruvasagar joined08:54
← Blahhh left08:55
→ wrouesnel1 joined08:57
EugeneKay o.O08:57
← wrouesnel left08:58
← rasto left08:58
iveqy cbreak: what has that have to do with it?08:59
vedranm iveqy: do you have a link on that?09:00
← donmichelangelo left09:00
cbreak maybe it'd be easier if git had a bit more code reuse, only one spot has to be changed :)09:00
→ donmichelangelo joined09:00
iveqy vedranm: link on what?09:00
vedranm switch to libgit209:00
iveqy there's no really decision, just hints I pick up from core developers in09:01
http://thread.gmane.org/gmane.comp.version-control.git/226751/focus=22675209:01
and specially an other thread (hold on)09:01
→ skorgon joined09:02
← wrouesnel1 left09:02
← yacks left09:02
← vedant left09:02
iveqy http://thread.gmane.org/gmane.comp.version-control.git/226424/focus=22680109:03
that thread09:03
cbreak: probabaly. You still need to consider that git isn't the only program that uses the git database, so it can be a bit messy...09:03
afaik libgit2 and git can coexists with the same repo today09:04
→ wrouesnel joined09:04
cbreak as long as they use the same library it's ok09:04
iveqy cbreak: yeah, but they don't...09:05
cbreak ssh3 repos would then be incompatible with the normal ones of course, but that's to be expected09:05
← jceb left09:05
iveqy I guess it wouldn't be too hard to support both sha3 and sha1 repos09:05
vedranm iveqy: doesn't seems so explicit09:06
I wouldn't count on that09:06
grawity what's the problem with sha1? the way it's used by Git, I don't see it breaking any time soon09:06
iveqy git's codebase is pretty tightly bound to sha1 though...09:06
vedranm: it's not explicit and it's not said by junio. However libgit2 has already had a few (failing) attempts to reinvent git09:07
and git is not working towards replacing libgit209:07
← barneybook left09:07
→ jceb joined09:07
iveqy and since libgit2 has been supported by microsoft they probably will continue growing09:08
cbreak like everything else microsoft supports? :)09:09
→ thierryp joined09:09
cbreak (They also support Visual Source Safe I heard...)09:09
grawity: sha3 has a bigger number, so it must be better09:10
grawity well, VSS must be why they switched to Git09:10
cbreak grawity: also, there are potential theoretical weaknesses in sha109:10
← dhruvasagar left09:11
cbreak (not an issue for now, but it could be in the future, better to have an exit strategy and not get fucked up like all those moronic "OH, SHA1 for our password databases is enough"... sha1 was never good enough for that kind of hashing...)09:11
for now it's more than enough for git09:11
but who knows09:11
vedranm iveqy: I see09:12
well, it will be obvious within the coming months09:12
who knows09:12
iveqy what I'm trying to say is that libgit2 seems to be a stable mature and usefull project09:12
grawity the security risks with password databases are different09:12
vedranm iveqy: absolutely09:12
iveqy vedranm: well maybe.09:13
cbreak most GUI tools seem to use plumbing09:13
iveqy I think it's kind of sad to have libgit2. git would be better if all efforts where focused in one place09:13
grawity I think having multiple implementations is good09:14
iveqy cbreak: on linux yes, but not on windows. More and more gui tools switch to jgit or libgit209:14
← thierryp left09:14
→ ntus1017 joined09:14
iveqy grawity: why's that? I'm curious09:14
grawity then you don't accidentally end up with something that's impossible to re-implement09:14
iveqy cbreak: git doesn't play well on windows because of it's perl bindings09:14
grawity: that's true. But you get a problem if the "reference implementation" isn't the best implementation anymore09:15
cbreak multiple implementations is good so people don't rely on implementation defined behavior09:15
← jceb left09:15
iveqy let's say that libgit2 is better than git, and the starts to add new features09:15
cbreak and instead expect the behavior as it is standardized09:15
grawity yes, what cbreak said09:16
EugeneKay The repo format and the git:// protocol is what matters09:16
iveqy true09:16
→ jceb joined09:16
iveqy however then git needs to have a standard. The only standard today is a reference implementation09:16
EugeneKay The behaviour of any implementation/program that uses this is really irrelevent09:16
cbreak it's like C++ vs some proprietary language09:17
if there's just one implementation, then it is very hard to write portable code09:17
EugeneKay iveqy: man gitrepository-layout09:17
gitinfo iveqy: the gitrepository-layout manpage is available at http://jk.gs/gitrepository-layout.html09:17
EugeneKay I'm unsure where git:// is documented09:17
cbreak I think I saw a huge text file in git.git once09:18
← jaequery left09:18
→ danicampa901 joined09:19
iveqy EugeneKay: EugeneKay thanks. And there seems to be somethings missing in that... Look at the "NOTE"-part in the beggining09:19
EugeneKay There are already several alternate git implementations, have been for years. git.git being the Official one is pretty new as of 1.7(I think?)09:19
iveqy those modules will be put in .git/modules/[path to submodule]. And that isn't mentioned09:19
EugeneKay: in what way did git.git became official in 1.7? I've missed that09:20
EugeneKay Patches welcome ;-)09:20
grawity patches to documentation – especially so09:20
iveqy EugeneKay: well, the thing is that it was me that implemented the module/-thing two years ago. So it's me that missed to patch that man-page anyway... :)09:21
EugeneKay If I read my history right, pre-1.7 the porcelain commonly known as `git` was known as `tig`09:21
grawity It was cogito, I think09:21
tig is the ncurses UI09:22
EugeneKay That one. tig is the ncurses... yeah09:22
And it was 1.509:22
http://git.or.cz/cogito/09:22
grawity (hmm, is it normal that the 'next' branch cannot always be fast-forwarded?)09:22
← curvedmark left09:23
→ vedant joined09:23
iveqy http://git.or.cz/cogito/ found it, thanks09:23
oh EugeneKay was faster =)09:23
grawity: no... but it could happen if something was removed from it, right?09:23
removing things from next happens regulairly, but not very often09:24
EugeneKay I would welcome a better CLI than `git`. Anybody who's read the checkout man page knows how terrible it is.09:24
grawity (I'm wondering about 'next' on git.git specifically)09:24
← elixir left09:24
→ elixir joined09:25
→ hamburger2000 joined09:25
→ marhaban joined09:25
Nevik EugeneKay: how exactly would you have a new cli client be set up? i mean most of the git terminology is directly related to the git commands09:25
so if you split what `checkout` does now into other (possibly new) commands, that'll be confusion hell09:26
just as an example09:26
(i agree that it's anything but intuitive, but i just wanna see what your thoughts on it are)09:27
→ St_Marx joined09:27
→ marhaban1 joined09:27
→ Cromulent joined09:27
EugeneKay I would limit a command to doing one thing09:28
Anything involving branch switching? `git branch`09:28
`git branch switch master`09:28
Nevik sounds good to me09:28
EugeneKay I would have `git checkout` deal with checking out a file, defaulting to HEAD09:29
→ ThomasCompix joined09:29
EugeneKay But.... I'm not about to write it ;-)09:29
← catsup left09:29
→ arnihr joined09:29
Nevik pity :P09:29
→ catsup joined09:29
Nevik EugeneKay: write the documentation, let someone else implement it? :P09:29
EugeneKay I have other things to write first09:29
Considered it once. Got bored09:30
Nevik haha09:30
→ beawesomeinstead joined09:30
→ sostenuto joined09:31
← wwu_ left09:35
→ rednaks joined09:40
→ thierryp joined09:40
← babilen left09:41
→ Shusshu joined09:42
SeoZ → SeoZ-work[AWAY]09:43
← thierryp left09:44
← hyperair left09:48
→ koltroll joined09:48
← Shusshu left09:48
→ hyperair joined09:48
→ fluxdude joined09:48
← ej left09:49
← hyperair left09:51
→ dhruvasagar joined09:51
→ wwu_ joined09:52
← paul424 left09:53
← mechanicalduck left09:55
← Cromulent left09:56
→ aard_ joined09:56
→ setmeaway2 joined09:56
← dhruvasagar left09:59
→ zacts joined09:59
← jceb left10:00
→ mechanicalduck joined10:00
→ Anders_J joined10:02
→ jceb joined10:02
← alexandernst left10:03
← jceb left10:04
→ hyperair joined10:04
← elixir left10:05
→ alexandernst joined10:05
→ jceb joined10:05
→ elixir joined10:05
← mechanicalduck left10:05
→ shruggar joined10:07
← jceb left10:07
← hyperair left10:07
→ jceb joined10:07
← aartist left10:08
→ dhruvasagar joined10:08
← danicampa901 left10:08
→ m8 joined10:09
→ mishok13 joined10:09
→ thierryp joined10:11
← ntus1017 left10:11
← jceb left10:12
→ jceb joined10:13
→ babilen joined10:13
← dhruvasagar left10:13
← thierryp left10:15
→ eli joined10:15
→ nicedice joined10:15
→ hyperair joined10:18
← howlymowly left10:18
→ ej joined10:20
← rudak1 left10:21
→ rudak1 joined10:22
← St_Marx left10:25
→ DonRichie joined10:26
← jceb left10:26
← maslowbeer left10:27
grawity Hmm, do the cogito repos still exist anywhere? They're not on kernel.org anymore.10:27
→ jceb joined10:27
← nicedice left10:27
← sallesao left10:27
← ej left10:28
grawity Found cogito.git on repo.or.cz, no cogito-doc.git though10:30
→ ryez_ joined10:30
grawity Ah, found in a mirror10:30
iveqy grawity: https://www.kernel.org/pub/software/scm/cogito/10:30
grawity: and why would you even be interested in that?10:31
just curious10:31
grawity iveqy: Those dirs are empty except for 'RPMS'10:31
iveqy: and, just because10:31
I like taking a look at old stuff, and I hate it when software just disappears.10:31
iveqy oh sorry. At least there's src-rpms10:31
→ nicedice joined10:32
iveqy yeah, me too. One piece of code I use is only hosted as a debian package today, kind of scarry10:32
← chamunks left10:32
→ chamunks joined10:33
← hyperair left10:34
→ Balliad joined10:34
← darnir left10:37
← chamunks left10:38
← luc_ left10:39
→ jceb_ joined10:41
→ thierryp joined10:41
← kcj left10:41
← jceb_ left10:41
→ nafg joined10:44
→ the_cybersphinx joined10:44
← abstrusenick left10:45
← thierryp left10:45
→ Fandekasp joined10:46
→ FernandoBasso joined10:47
← nicedice left10:49
← jceb left10:50
→ aChatir joined10:53
→ jceb joined10:54
← jceb left10:54
← avinashrbhat left10:55
→ nicedice joined10:55
→ LuxuryMode joined10:57
← loops left10:59
← LuxuryMode left10:59
← Aprogas left11:01
→ Aprogas joined11:01
← elixir left11:04
→ jceb joined11:04
→ elixir joined11:04
→ avinashrbhat joined11:05
→ LuxuryMode joined11:05
← raisen left11:06
← LuxuryMode left11:06
→ Dave^| joined11:06
→ martijnHH joined11:07
→ LuxuryMode joined11:07
→ loops joined11:07
→ dhruvasagar joined11:11
→ barq joined11:11
→ yacks joined11:11
→ hyperair joined11:11
→ thierryp joined11:12
→ mklappstuhl joined11:13
barq Is there a channel for git development?11:14
EugeneKay Mailing list.11:15
barq I'm on that but have no idea where to start, there's so many messages flooding my inbox.11:15
EugeneKay Yup. Mailing lists suck.11:16
← thierryp left11:16
← echo-area left11:17
iveqy barq: the list is pretty noisy yes. But usually works well otherwise11:18
barq: can I help you with anything?11:18
← nerdfiles left11:18
← danmackay left11:19
← SamB left11:20
→ SamB joined11:20
→ ej joined11:20
← Balliad left11:20
→ danmackay joined11:22
charon barq: #git-devel11:23
but that's just if you want generic advice and such, for in-depth questions the list is much better... and also, many of the area experts aren't on IRC11:23
← ej left11:25
→ Ziber joined11:27
→ ozette joined11:27
← jceb left11:30
← longbeach left11:33
← Dave^| left11:35
→ leeN joined11:36
← martijnHH left11:38
← ryez_ left11:38
← mklappstuhl left11:39
→ jceb joined11:39
esc_ what would happen if there is a SHA collision in a Repo11:40
← danmackay left11:40
→ Droolio joined11:42
→ danmackay joined11:42
→ danicampa901 joined11:42
→ thierryp joined11:42
grawity it's unlikely enough that nobody really cares what would happen11:43
→ murr4y joined11:43
grawity but if you had two files which, with the Git object header, had identical SHAs, then git would just never add the second file11:44
← leeN left11:44
← babilen left11:45
grawity so when you cloned the repo, it'd have file A's contents in both file A and B11:45
but given that both files must also be of the same length (part of the header), I think collissions are unlikely enough11:45
→ leeN joined11:47
← thierryp left11:47
→ jnavila joined11:49
→ system64 joined11:50
→ laissez-f joined11:51
→ gusnan joined11:51
→ mechanicalduck joined11:51
→ w0ng joined11:51
← gusnan left11:52
→ exoon joined11:52
← BiggFREE left11:55
ras0ir → akgenclik11:56
→ aendrew joined11:57
← jayfk left11:57
→ ntus1017 joined11:59
← yuwang left12:00
← vedant left12:03
← system64 left12:04
← pac1 left12:05
← yacks left12:07
→ Noldorin joined12:08
→ mixel joined12:09
EugeneKay If you do find a SHA1 collision, be sure to tell all of the security mailing lists12:11
Because there are no (publically) known ones yet.12:11
← marhaban1 left12:13
← Sigma left12:13
→ thierryp joined12:13
specing Seriously?12:15
← LuxuryMode left12:16
→ LuxuryMode joined12:17
→ threesome joined12:17
→ marhaban1 joined12:18
← Jahm left12:18
→ drizzd joined12:18
← thierryp left12:18
← jokajak` left12:19
→ jgh joined12:19
→ ej joined12:21
→ m0viefreak joined12:23
← d2dchat left12:23
→ codertux joined12:25
← ej left12:26
→ Sigma joined12:26
← w0ng left12:27
→ hellopat joined12:28
← dhruvasagar left12:29
→ Jahm joined12:29
→ ej joined12:30
← jceb left12:31
EugeneKay Yup.12:31
← aChatir left12:33
EugeneKay An attack has been found that reduces the effort to 2^63 invocations(rather than the brute-force 2^80), and a collisison has been demonstrated on a 53-round variant.12:33
← leeN left12:34
→ sr78ger joined12:34
← sr78ger left12:35
→ chamunks joined12:35
← marhaban1 left12:35
→ leeN joined12:35
→ sr78ger joined12:35
← PaulCapestany left12:36
← trusktr left12:37
← Milossh left12:38
→ AdrienBrault joined12:39
cbreak so far, sha1 has some theoretic weaknesses12:41
sha3 is completely different and does not suffer from the same weaknesses (but probably other ones)12:41
sha2 is as far as I know an improvement on sha112:41
→ ddr2 joined12:41
cbreak so it might have similar shortcommings12:42
← LuxuryMode left12:43
→ Milossh joined12:43
→ jcao219 joined12:44
→ thierryp joined12:44
← arand left12:44
← leeN left12:46
← AdrienBrault left12:47
→ AdrienBrault joined12:48
← thierryp left12:49
→ ThomasLocke joined12:50
→ chrob joined12:50
EugeneKay It's longer, and no ueful attacks have been brought agaisnt it12:50
← alexander__b left12:51
→ alexander__b joined12:51
→ lonewulf` joined12:51
→ babilen joined12:54
→ kkthecoder joined12:55
→ SergeyPenkin joined12:57
→ simesy joined13:00
← lonewulf` left13:01
→ marhaban1 joined13:01
→ pastime joined13:02
← odin_ left13:03
← ntus1017 left13:03
← krion left13:04
← aendrew left13:04
← danicampa901 left13:07
→ LuxuryMode joined13:08
← virrus_home left13:09
→ virrus_home joined13:10
→ Raging_Hog joined13:12
→ ntus1017 joined13:12
→ jceb joined13:14
→ thierryp joined13:15
← pastime left13:15
→ odin_ joined13:17
← OOPMan left13:17
← thierryp left13:20
→ nsaibot joined13:20
← wwu_ left13:21
← pandeiro` left13:21
→ vitor_ joined13:23
grawity It's longer, but I seem to remember it having some disadvantages even compared to SHA1.13:24
nsaibot any ideas as to how to set the hunk header to a prebuilt formats (let's say cpp) in the .gitattributes?13:26
← vitor__ left13:26
← jceb left13:27
→ rasto joined13:28
→ dhruvasagar joined13:28
→ howlymowly joined13:28
zizzfizzix_ → zizzfizzix13:28
→ mechanicalduck_ joined13:30
→ Cromulent joined13:30
← mechanicalduck left13:32
mechanicalduck_ → mechanicalduck13:33
← ntus1017 left13:33
← dhruvasagar left13:33
→ tkrugg joined13:36
← tkrugg left13:36
← barq left13:36
← fobban left13:38
← pantsman left13:40
→ pastime joined13:40
→ _BJfreeman joined13:40
→ R1KRDO joined13:41
← Cromulent left13:41
_BJfreeman → BJfreeman13:41
→ Cykey joined13:41
→ relipse joined13:42
← segher left13:43
← kpreid left13:44
→ noxd joined13:45
→ segher joined13:45
→ kpreid joined13:45
akgenclik → ras0ir13:45
← kpreid left13:45
→ szx joined13:45
→ kpreid joined13:46
← candlejackson left13:46
→ thierryp joined13:46
→ konundra joined13:47
→ mechanicalduck_ joined13:48
→ mateusz__ joined13:48
→ gasbakid_ joined13:48
→ candlejackson joined13:48
→ motto joined13:50
← gasbakid left13:50
← mechanicalduck left13:50
← thierryp left13:51
→ mckendricks joined13:51
← m8 left13:52
← beawesomeinstead left13:52
→ beawesomeinstead joined13:52
← szx left13:52
← charliesome left13:53
→ carlos-huang joined13:53
→ m8 joined13:55
← motto left13:55
→ mribeirodantas joined13:56
→ TomyLobo joined13:56
→ penk joined13:56
→ gdelfino joined13:56
→ m0ltar joined13:56
← j416 left13:56
gdelfino Is public push access to docrails closed? I am getting "ERROR: Permission to rails/docrails.git denied to gdelfino"13:57
→ j416 joined13:57
cmn gdelfino: you should ask them13:58
gdelfino Sorry, I mistakenly asked here.13:58
→ vigintas joined13:59
← babilen left14:00
← nsaibot left14:00
→ gusnan joined14:00
→ danicampa901 joined14:02
→ sgambino joined14:02
→ vedant joined14:03
← elixir left14:04
← homer91 left14:04
→ lindenle joined14:05
→ rath joined14:05
← wrouesnel left14:05
← rudak1 left14:05
→ chrisshattuck joined14:09
← simesy left14:11
← BJfreeman left14:11
→ jelmer joined14:12
→ Dougie187 joined14:12
→ mechanicalduck joined14:13
→ thierryp joined14:14
→ a-l-e joined14:15
→ Thanatermesis joined14:16
→ _BJfreeman joined14:16
→ dsdeiz joined14:16
Thanatermesis i had a power outage and now I can't use my git repository, when i enter on the dir it shows me an error about "object file is empty" | "fatal: loose object (stored in .git/objects/...) is corrupt14:16
_BJfreeman → BJfreeman14:17
← mechanicalduck_ left14:17
→ freestyl3r joined14:17
Thanatermesis what i can do now ? (i want to recover the usability of my repository, of course, but i want to know also "what is lose here"14:17
← thierryp left14:19
→ JEEBcz joined14:19
← FernandoBasso left14:20
← ej left14:21
→ ej joined14:21
← JEEB left14:22
→ kukks joined14:23
← kukks left14:23
→ kukks joined14:23
iveqy Thanatermesis: I'm not sure about this. But I would suggest doing a backup of your repo and then look into git fsck14:23
→ system64 joined14:23
Thanatermesis seems like filesystem corruption is not a very known science...14:24
→ leeN joined14:24
jgh assume you have fsck's the14:24
Thanatermesis is still surprised how we can have filesystem corruptions by a simple power outage these years with the lot of development and improvments made14:24
jgh underlying filesystem14:24
iveqy Thanatermesis: well, I'd the luxury of never having that problem. and if I do, I just pulled my backup.14:24
← leeN left14:24
iveqy Thanatermesis: oh it's not. Linux doesn't write to the disc at once, but does delayed writing. It's a performance thing14:25
Thanatermesis iveqy, its not so easy, i have backups too, but how i can know "which files are lose/corrupted" ? i just discovered the failure on one of my git repos, because i have tried to use it14:25
iveqy Thanatermesis: you could have instants writes to the disc, that would help you with your power outages, but that would slow you down14:25
← Dougie187 left14:25
iveqy Thanatermesis: oh, that's why git fsck is for, try it14:26
a nice thing with git is that you can tell which files that are corrupted. with svn you can't14:26
Thanatermesis ok ill make a look to git fsck then, btw is there any channel more specific for filesystem problems ?14:27
from line 730 it means that it has removed files on my system? http://paste.debian.net/9335/14:28
← Exio left14:28
→ Exio joined14:29
← donmichelangelo left14:30
→ leeN joined14:30
→ donmichelangelo joined14:30
→ CocoStorm joined14:32
iveqy Thanatermesis: you get better answers in #linux (or maybe something like #ext4)14:33
CocoStorm i tried to create a new rep using git init and it was fine but when i log onto github it's not there14:33
and when i try git push origin master it complains that origin is not a rep14:33
grawity Did you create the repo on Github too?>14:33
Thanatermesis iveqy, ok thx14:33
CocoStorm grawity, nope14:33
grawity Then of course it's not on Github14:33
CocoStorm grawity, can i do it fom command line14:33
grawity, i see14:33
so i can't do it all from terminal?14:34
grawity the repo on your computer and the repo on Github (and the repo on someone else's computer) are completely independent14:34
charon CocoStorm: there's a 'hub' utility that lets you do github operations from the command line14:34
grawity it just happens that you can push and pull between them14:34
CocoStorm grawity, i see and when i do git push origin master does it know which rep to push it to?14:34
sorry for the noob questions14:34
grawity CocoStorm: you tell it14:34
CocoStorm grawity, how?14:34
charon, oh i see, i'll look into it14:34
grawity CocoStorm: when you do `git remote add origin git://github.com/foo/bar`14:34
CocoStorm oh i see14:35
i shuold do that then14:35
grawity CocoStorm: (or when you `git clone` an existing repo – then the clone address automatically gets added)14:35
CocoStorm makes sense, thanks guys!14:35
thanks grawity14:35
will try it out14:35
grawity CocoStorm: after you create the repo on Github, it gives you short instructions on pushing from an existing local repo14:35
or, as charon said, `hub create` with the `hub` script14:35
← lindenle left14:36
CocoStorm brilliant, it worked thanks :)14:37
and yeh i will try and see how the hub script works14:37
← candlejackson left14:37
→ candlejackson joined14:39
← jcao219 left14:39
← JEEBcz left14:42
→ aweorit joined14:43
← relipse left14:44
← mixel left14:45
← carlos-huang left14:45
→ JEEB joined14:46
← bret left14:47
→ lindenle joined14:48
← CocoStorm left14:49
← sgambino left14:49
→ avtobiff joined14:50
← BJfreeman left14:52
← candlejackson left14:52
← hellopat left14:53
← system64 left14:54
→ _BJfreeman joined14:54
→ candlejackson joined14:54
← Fandekasp left14:55
_BJfreeman → BJfreeman14:55
→ system64 joined14:55
← lindenle left14:56
← takezawa left14:57
→ neutrino joined14:57
neutrino hi14:57
gitinfo neutrino: 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.14:57
← aard_ left14:58
→ CocoStorm joined14:58
← rednaks left14:59
neutrino i have got a problem. I wanted to check out an earlier commit, and used gitg in which I selected the commit and did "reset master branch to here". Now I don't know how to get my local commits back. How can I do this?14:59
→ aard_ joined14:59
neutrino sorry it was gitk which I used14:59
← CocoStorm left15:00
neutrino i'm afraid that my work is gone (I haven't pushed my commits yet..)15:00
→ kfogel joined15:01
← ozette left15:02
Nevik neutrino: is gitk still open and showing the later commits?15:03
and dont worry, your commits arent gone15:03
→ rednaks joined15:03
neutrino Nevik: it's not15:04
→ jaequery joined15:04
→ bret joined15:04
neutrino Nevik: i'm reading stuff on the internet, they say to do git reflog, and then merge to the last hash15:04
→ guiambros joined15:05
Nevik neutrino: did you make any commits after resetting?15:05
neutrino nope15:05
my reflog starts like this:15:05
985c2ae HEAD@{0}: reset: moving to 985c2ae73f28b4f106d745d51b957a7e1d6488cb15:05
938abe2 HEAD@{1}: commit: simplify ServerRunner15:05
the "simplify ServerRunner" thing is the latest commit I had before resetting.15:06
← leeN left15:06
← LuxuryMode left15:06
Nevik neutrino: are there any uncommitted changes that you care about?15:06
neutrino nope15:06
Nevik then just `git reset --hard 938abe2`15:06
neutrino why not merge? what's the difference?15:06
→ nbouscal joined15:07
Nevik in this case, there is none. you can merge if you feel safer with it15:07
neutrino i don't :)15:07
→ natechan joined15:07
Nevik it is, in fact, a bit safer, if you had made any commits or uncommitted changes15:07
but since i asked you about those, reset wont lose anything15:07
neutrino thank you ery much15:07
git is so scary15:08
can you recommend a UI such as gitk that allows me to check out earlier commits?15:08
i guess resetting isn't the right way to do this, right?15:08
← iveqy left15:08
← Sigma left15:08
neutrino i want to browse earlier versions to see when a specific bug started happening15:08
Nevik neutrino: it's not. `git checkout` is the right way -- as the name suggests15:08
neutrino: i dont really have a lot of experience with guis since i only use the cli15:09
i only use gitk for history browsing15:09
neutrino yeah same here15:09
→ jayfk joined15:09
neutrino but once you're in the history it might be nice to be able to just click something, rather than re-type hashes15:09
Nevik neutrino: if you have many revisions to comb through, look at man git bisect15:09
gitinfo neutrino: the git-bisect manpage is available at http://jk.gs/git-bisect.html15:09
neutrino will browse15:09
neutrino i know of bisect, thanks15:09
← natechan left15:09
neutrino it's not so much though15:09
Nevik ah, well15:09
neutrino thanks Nevik15:10
:)15:10
Nevik neutrino: if you want to go into the past, rather than forward from older commits, you can `git checkout HEAD^` on every commit to get its parent15:10
→ tplaner joined15:10
Nevik the only place to pay attention there is when youre on a merge commit, since it'll have two parents15:10
← gusnan left15:11
neutrino mhm15:11
that's a good tip, thank you15:12
← mechanicalduck left15:12
→ mechanicalduck_ joined15:12
← system64 left15:13
→ fobban joined15:17
← m8 left15:17
→ rgr joined15:18
← CodeShark left15:18
→ LuxuryMode joined15:19
→ w0ng joined15:21
← jayfk left15:22
→ jokajak joined15:23
nbouscal neutrino: SourceTree seems to be a decent gui. I stick to the cli myself but my coworkers seem happy with it15:24
neutrino gotcha15:24
i'd love something in curses tbh15:24
but you just can't do a nice branch graph in curses.15:24
nbouscal have you looked at tig?15:24
specing tig is nice15:24
neutrino or rather, you can, but people haven't, because they're scared of ansi.15:24
nope, what does tig give me?15:24
nbouscal history in ncurses, with a few other things thrown in15:25
grawity a nice branch graph among other things15:25
nbouscal not sure if you can checkout from within it though15:25
grawity sure, bind a key15:25
← BJfreeman left15:25
neutrino looking15:25
nbouscal oh with the external command thing? cool. i haven't played with it much, i have custom history commands that make it mostly redundant for me15:26
neutrino right, i have it installed already15:26
nice15:26
does tig have a thing to do partial commits? i.e. select hunks/lines?15:26
i have to use git gui for that (git interactive commit is too burdensome)15:27
grawity well, there's `git add -p`15:27
nbouscal i almost never use interactive add, but i use patch add all the time15:27
→ TheDracle joined15:27
→ tanath joined15:27
neutrino nah i was asking about tig specifically15:27
grawity but yes, tig has that.15:27
neutrino nice, how do i do this?15:27
grawity 'S' → 'u'/'1'15:27
neutrino looking15:27
→ _BJfreeman joined15:30
← AdrienBrault left15:31
_BJfreeman → BJfreeman\15:31
BJfreeman\ → BJfreeman15:32
→ carlos-huang joined15:34
← kkthecoder left15:35
→ Goplat joined15:38
← Nevik left15:39
← virrus_home left15:40
← needlz left15:40
→ virrus_home joined15:40
→ Nevik joined15:40
→ needlz joined15:41
← Davey left15:41
← TheDracle left15:42
← nicedice left15:43
→ nicedice joined15:44
← rath left15:44
← Adlai left15:45
→ elixir joined15:45
neutrino grawity: u only lets me stage the files as a whole. It doesn't let me select which lines to stage.15:45
grawity Press Enter instead15:45
→ jhoepken joined15:45
neutrino looking15:46
← marhaban1 left15:46
neutrino great!15:47
that's exactly what i've been looking for15:47
thank you grawity15:47
are you the tig developer?15:47
grawity No.15:47
neutrino hmm do you know how long this feature has been in tig for?15:47
is it new?15:47
i've had a look at tig before but either missed it or it wasn't there, that's why i started using git gui15:48
Nevik neutrino: did you read the documentation at the time?15:49
→ rpg joined15:50
→ marhaban1 joined15:51
← marhaban left15:51
→ leeN joined15:51
→ jnavila_ joined15:52
→ chrisirc joined15:53
← ddr2 left15:53
chrisirc What's the best (most rememberable) way to change push settings to push to another branch?15:54
i.e. my workflow is:15:54
On machine A (server): mkdir foo; cd foo; git init; hackaway; git commit;15:54
← vedant left15:55
→ ntus1017 joined15:55
chrisirc On machine B (desktop): git clone A:foo/.git; hackaway; git commit; <missing>; git push15:55
charon you should use a !bare repository for this15:55
gitinfo an explanation of bare and non-bare repositories (and why pushing to a non-bare one causes problems) can be found here: http://bare-vs-nonbare.gitrecipes.de/15:55
← fluxdude left15:55
chrisirc I want a <missing> step so that I can log back into A, then "git merge B_master" or so.15:55
← kfogel left15:55
offby1 chrisirc: actually I don't see anything wrong with what you're doing -- even if <missing> is empty15:55
chrisirc charon, I know about bare repositories, but this workflow just suits me fine.15:55
neutrino Nevik: i don't remember. Most likely.15:56
chrisirc offby1, push will push to master there, in a non-bare rep15:56
grawity git push origin master:B_master15:56
→ mklappstuhl joined15:56
← jnavila_ left15:56
grawity hmm, I wonder if there's something like a default refspec for pushing15:56
charon chrisirc: in that case, you can configure remote.origin.push = refs/heads/*:refs/remotes/B/*15:56
grawity oh, cool15:56
chrisirc charon, yep, I know, but I always forget the details there. And I don't want to always figure it out again...15:57
charon write it on a piece of paper and glue it to your screen15:57
chrisirc Maybe I'm just too stupid to understand the concept of those paths (which is local which is remote, etc.)15:57
Hm. My way to write that stuff down is to write a script to do it.15:57
But I thought maybe I don't have to.15:58
charon on a more serious note, it's not that hard; these fetch/push specs are of the form src:dst and take all the refs matching the src glob on the src side, and push them to the dst side with the * replaced by what matched on the src side15:58
s/push/fetch or push/15:58
chrisirc Is src vs dst consistent in some way?15:58
charon huh?15:58
it's always src:dst15:58
chrisirc i.e. "git push foo:bar" I also always forget which is which.15:58
offby1 chrisirc: I completely identify with your confusion. But this is one case where I think it's best to just learn the details.15:59
(as opposed to saving yourself from having to memorize them by hiding them in a script)15:59
charon e.g. the default fetch spec for a remote called origin is refs/heads/*:refs/remotes/origin/* so that you get the remote (source for fetch) local branches mapped to the local (dest for fetch) remote namespace origin/*15:59
chrisirc: the added bonus once you know what they do is that you can do mass branch copies by saying 'git fetch . foo/*:bar/*' etc16:00
(the . here means: the same repository is also the source)16:00
chrisirc alright, going to get a coffee to mod up my brain cells, then do it that way.16:00
← devoldmx left16:03
hakunin_ → hakunin16:05
← j416 left16:06
→ j416 joined16:06
→ devoldmx joined16:06
→ fluxdude joined16:07
→ HerrK joined16:08
→ kfogel joined16:08
→ vedant joined16:10
offby1 coffee++16:11
→ aartist joined16:11
← jaequery left16:12
charon is there a utility like cat, but that prefixes the filename to each line? it seems a bit lame to use 'grep -H -e "" <files>' for that purpose16:13
alfplayer Hi. Can I clone the history without any content, without blobs and without checking out, only to see the history? I want to minimize the amount of downloaded data.16:16
charon alfplayer: no.16:16
← mklappstuhl left16:16
alfplayer Thanks.16:16
→ i42n joined16:16
charon well, not much help, sadly16:16
← i42n left16:16
charon there's no deep reason that prevents this from happening, it's just a feature that nobody cares for16:16
→ i42n joined16:17
charon (as opposed to e.g. restartable clones, where there *is* a reason)16:17
← vigintas left16:17
← devoldmx left16:17
alfplayer Someone is using a git repo to store artwork only. I think he shouldn't have.16:18
→ devoldmx joined16:18
→ asteros joined16:18
→ franck joined16:18
franck → Guest1274016:18
← Guest12740 left16:19
→ Guest12740 joined16:19
→ rudak joined16:22
← carlos-huang left16:22
→ Cromulent joined16:22
→ lindenle joined16:23
← Guest12740 left16:25
→ Liir joined16:26
→ tjbiddle joined16:26
← leeN left16:26
→ Sigma joined16:27
→ leeN joined16:27
jgh charon: awk '{print FILENAME ": " $0}' foo16:29
→ tjbiddle_ joined16:31
← tjbiddle left16:31
tjbiddle_ → tjbiddle16:31
toky_ → toky16:33
→ OOPMan joined16:35
→ sgambino joined16:37
→ amcsi joined16:38
← Cromulent left16:42
→ babilen joined16:43
→ AdrienBrault joined16:43
→ thiago joined16:44
→ oriba joined16:45
← dsdeiz left16:47
← AdrienBrault left16:47
→ BSaboia joined16:49
→ sgambino_ joined16:49
← m0ltar left16:50
← LuxuryMode left16:50
← sgambino left16:51
sgambino_ → sgambino16:51
Nevik jgh: awk --ward16:51
charon jgh: heh. that's cheating! ;-)16:51
i wanted to retort that perl surely has some $[@#*%&!] magic for FILENAME, but it turns out it's called $ARGV. how boring.16:52
offby1 alfplayer: randallschwarz often says: if you don't plan to merge these files, they shouldn't be in git16:53
alfplayer offby1: Interesting. Thanks.16:53
offby1 hey, 'awk' is the granddaddy of the good scripting langauges16:53
alfplayer I guess he was tempted because the rest of our repos are all git.16:53
→ sgambino_ joined16:54
offby1 it's not a crime or anything, but if you're not going to merge, then there might be better tools around to do whatever it is you need to do. (rsync and friends, for example)16:55
← hyperair left16:55
← Dead_Body left16:55
← sgambino left16:56
sgambino_ → sgambino16:56
← konundra left16:56
← noxd left16:57
→ barneybook joined16:57
← kortti_ left16:59
→ sajan joined17:00
→ tmcmahon joined17:01
← gittish left17:02
← rgr left17:02
→ jkubecki joined17:02
← jhoepken left17:03
→ spaceonline joined17:03
→ gittish joined17:04
← HerrK left17:05
→ antoine__ joined17:05
→ lresende joined17:05
← cagedwisdom left17:05
→ system64 joined17:06
gitinfo set mode: +v17:06
← Z6 left17:06
← asteros left17:10
→ kortti joined17:10
← system64 left17:10
→ asteros joined17:10
→ sq-one joined17:12
offby1 ~.17:12
← i42n left17:12
offby1 heh17:12
← hamburger2000 left17:13
→ conan_ joined17:14
→ SBoudrias joined17:14
→ yacks joined17:15
← j416 left17:15
→ j416 joined17:15
→ system64 joined17:15
← j416 left17:16
→ j416 joined17:16
Nevik lol offby117:18
→ Dougie187 joined17:18
Nevik everyone has ls'd in irc before :P17:18
Dougie187 Have they all `rm -rf *`'d too? :P17:18
← Raging_Hog left17:18
Nevik Dougie187: better in irc than in a shell :P17:19
Dougie187 true.17:19
Nevik well, usually that command doesnt break much on a system-level17:19
it might lose you some work, but at least it doesnt wreck your OS17:20
unlike `sudo rm -rf /` haha17:20
← johnw left17:21
→ hyperair joined17:21
← rednaks left17:22
← Liir left17:22
Dougie187 lol17:23
← SergeyPenkin left17:24
← virrus_home left17:24
→ virrus_home joined17:25
→ mechanicalduck joined17:26
← angelsl left17:26
→ mixel joined17:27
→ aendrew joined17:28
← mechanicalduck_ left17:30
→ thomas_adam joined17:31
→ Nicholas_Rishel joined17:32
→ jayfk joined17:32
← Sigma left17:36
← pastime left17:36
← tmcmahon left17:37
charon Nevik: too bad that (barring backups, which you should have of both) restoring "some work" can mean considerable effort, while rebuilding an OS from scratch takes less than an hour if you use any of the modern binary distributions17:39
← jayfk left17:39
Nevik well, that depends on how custom your OS setup is17:40
installing all the applications you need to work might take well over an hour17:40
charon: but of course, both situations can turn out bad :P17:40
→ LuxuryMode joined17:41
Nevik and also, "delete everything" includes "delete all my work" -- so that's my trump here ;)17:41
← jkubecki left17:42
← j416 left17:42
→ j416 joined17:43
→ paul424 joined17:43
← j416 left17:44
→ j416 joined17:44
← BSaboia left17:45
← mckendricks left17:47
← virrus_home left17:48
→ virrus_home joined17:49
→ johnw joined17:49
← conan_ left17:49
← LuxuryMode left17:51
→ vadelius joined17:51
→ jayfk joined17:53
← sgambino left17:53
← aweorit left17:53
→ LuxuryMode joined17:53
→ pen joined17:54
→ pastime joined17:55
← vadelius left17:56
→ kaptoxic joined17:56
← LuxuryMode left17:57
→ Adlai joined17:57
← lindenle left17:58
→ bvi joined18:00
← spaceonline left18:00
← johnw left18:03
← Kadmos left18:03
→ Azis joined18:03
→ rednaks joined18:04
→ Sigma joined18:05
→ oriba_ joined18:06
← devoldmx left18:09
→ jaequery joined18:09
→ devoldmx joined18:10
← oriba left18:10
← aartist left18:10
← sq-one left18:10
← jaequery left18:11
→ vigintas joined18:11
→ jceb joined18:11
← _iron left18:12
← elixir left18:13
→ jaequery joined18:13
← devoldmx left18:13
→ jkubecki joined18:16
← nioncode left18:17
→ nioncode joined18:17
← j416 left18:19
→ j416 joined18:20
← leeN left18:20
→ leeN joined18:20
← jkubecki left18:21
← leeN left18:21
← aendrew left18:21
← exoon left18:22
→ leeN joined18:22
→ swiety joined18:22
← swiety left18:23
→ johnw joined18:23
→ noxd joined18:23
→ Bass10 joined18:23
→ swiety joined18:23
→ Guest67291 joined18:24
← DrBenway left18:27
→ i42n joined18:28
→ aendrew joined18:29
legumbre_ → legumbre18:29
← johnw left18:30
← Sigma left18:31
← asteros left18:31
→ Vile1 joined18:32
← Vile1 left18:32
adhawkins-away → adhawkins18:33
← Vile left18:34
← SJr left18:34
→ SJr joined18:34
← i42n left18:35
← SJr left18:35
adhawkins → adhawkins-away18:35
→ SJr joined18:36
→ thierryp joined18:36
← SJr left18:37
← vmiklos left18:37
→ vmiklos joined18:37
→ SJr joined18:38
→ phipes joined18:38
→ darnir joined18:42
← rudak left18:45
→ rudak joined18:45
← aendrew left18:47
← thierryp left18:48
← bvi left18:49
→ AWinterman joined18:49
→ BSaboia joined18:50
→ thierryp joined18:51
→ jkubecki joined18:52
→ i42n joined18:53
← i42n left18:53
→ torbjorn joined18:54
← jkubecki left18:57
← parasight left18:58
← AWinterman left18:59
← torbjorn left19:00
← sr78ger left19:01
← Ivoz left19:01
← Yakko left19:02
← mateusz__ left19:03
→ aendrew joined19:03
→ hellopat joined19:04
→ donnib joined19:05
← Azis left19:05
→ Kadmos joined19:06
→ taichunmin joined19:07
→ akurilin joined19:09
← mixel left19:09
→ Vile joined19:09
→ torbjorn joined19:09
Thanatermesis i have an empty/corrupt object in my git repo, which makes the repo entirely unusable, how i can fix the problem ? (i had a forced shutdown yesterday)19:09
_ikke_ Thanatermesis: Do you have recent backup or pushed recently to another repostory?19:10
Thanatermesis _ikke_, no, i use it only for local history (is my dir of tools/scripts)19:10
_ikke_ Thanatermesis: Were you doing any git operation at the point of shutdown?19:10
esto Thanatermesis: how is elive? :O19:11
← freestyl3r left19:12
loops Thanatermesis: did the error print the ID of the object? if not you can find it with git fsck -full19:12
Thanatermesis _ikke_, probably, this repo is in fact non-interactive, i have a script that does commits each 5 minutes for save an history of the work19:12
esto, very good thx :), working on the next version19:12
→ harobed_ joined19:12
→ m0ltar joined19:13
Thanatermesis loops, yes, there's 2 objects matched with the error19:13
Nevik Thanatermesis: you should mention that youve already asked this in here earlier and already did a git-fsck19:13
and re-post that link from earlier19:13
→ freestyl3r joined19:13
Thanatermesis yes, i have run fsck and it prints me these 2 objects, but i dont see any option to "fix" the problem in the man of fsck19:13
cmn it can't be fixed19:14
→ AdrienBrault joined19:14
cmn you need to get the repo from backup19:14
loops Thanatermesis: and then delete .git/objects/XX/Y{38} where XX is first 2 digits of sha1 and Y{38} is all remaining digits19:14
cmn, it can probably be fixed19:14
antoine__ → apelisse19:14
cmn only by getting the objects from backup19:14
→ Davey joined19:14
loops cmn, naw, they were just being created19:14
cmn, might lose most recent commit19:15
but nothing else19:15
cmn that's something you're just assuming out of the blue19:15
Thanatermesis what i wonder is what will happen to the files (or how they will act) if i delete these objects, mmh..19:15
loops cmn, no more of an assumption than saying it can't be fixed19:15
cmn, is worth a try19:15
Thanatermesis in fact the history is not important on this case (just a list of 5-minute-each commits)19:15
cmn except it's not an assumption19:15
the only way to repair this is to get the objects19:15
← system64 left19:15
loops cmn, shrug.. no reason to be so pedantic19:16
Nevik loops: this is #git19:16
pedantic is written in bold caps over the door to this room19:16
Thanatermesis maybe you mean to rewrite the history in order to be "correct up to now" ?19:16
cmn loops: no reason to be cavalier about the truth19:16
loops cmn, lol.. okay bud19:16
heh Nevik19:16
→ mixel joined19:16
cmn quite19:17
Thanatermesis well, what i do then ? delete the corrupted objects ?19:17
Thanatermesis feels like a good betatester19:17
cmn that won't help solve anything19:17
→ AaronMT joined19:17
_ikke_ Thanatermesis: first make a backup of the repo19:17
← aendrew left19:17
cmn if they are from the latest commt, you can re-write the index, assuming that's not broken and do somthing like commit --amend19:17
Thanatermesis already done :)19:17
loops Thanatermesis: listen to _ikke_19:17
Thanatermesis: then yeah, delete the corrupt objects19:17
Thanatermesis: should get you unwedged19:17
cmn but if you don't know what the objects refer to, then you won't be able to recreate them19:17
loops cmn, who cares19:18
← harobed_ left19:18
Nevik Thanatermesis: poke me when you're done with fixing this and/or before leaving; ive got a question for you but dont want to weigh in too much until youve got this dealt with19:18
Thanatermesis git status or gitk doesnt shows anything becuse of the error (well, only the error of the state is reported)19:18
→ johnw joined19:18
loops Thanatermesis: give it a go, let us know19:18
← hoschi left19:18
→ mklappstuhl joined19:18
Nevik Thanatermesis: lol you rhyme things19:18
← vedant left19:19
Thanatermesis well, objects removed19:19
← njbair left19:19
Thanatermesis now git status shows:19:19
fatal: bad object HEAD19:19
← lresende left19:20
_ikke_ Looks like HEAD points to something git can't read19:20
→ vigintasl joined19:20
→ harobed joined19:20
Thanatermesis cat .git/refs/heads/master -> a96eed29ad4483f5c3a28fd856a7420dbc2ba6b219:20
which is one of the ones that i have removed :)19:20
← harobed left19:20
→ harobed joined19:21
loops Thanatermesis: what does, git log master give ?19:21
ah... my bad, should have read19:21
Thanatermesis fatal: bad object master19:21
← harobed left19:21
← zacts left19:22
_ikke_ Thanatermesis: Then you should recreate master and point to something existing19:22
Thanatermesis: git reflog might help19:22
Thanatermesis _ikke_, "git reflog" only ?19:22
fatal: bad object HEAD19:22
_ikke_ Thanatermesis: It shows a list of things HEAD pointed to19:22
→ system64 joined19:22
_ikke_ ugh19:22
a bit less easy: .git/logs/...19:23
cmn loops: did you seriously just say "who cares" to data loss?19:23
loops yeah.. maybe just cat .git/logs/HEAD19:23
cmn, i'm done trying to explain it to you19:23
→ aendrew joined19:23
→ malkomalko joined19:23
← vigintas left19:23
→ lresende joined19:24
← Cykey left19:24
→ asteros joined19:24
Thanatermesis i have just noticed that the other object was not removed... doing it now19:25
→ SeptDePique joined19:25
→ Ivoz joined19:25
Thanatermesis cmn, on my exact case is not very important to maintain a perfect history, the really important thing is the last state of my codes19:25
← thierryp left19:26
← koltroll left19:26
→ whowantstolivef1 joined19:26
Thanatermesis one of the corrupted objects pointed to the last one, the other... idk, seems older19:26
← sostenuto left19:27
loops Thanatermesis: did you find a value for HEAD yet?19:27
→ koltroll joined19:27
loops either in .git/logs/HEAD or .git/logs/refs/heads/master19:27
← koltroll left19:27
Thanatermesis ah! this is interesting... i removed the second object, then running again "git fsck" shows me a new corrupted object19:27
then i remove it too, and git fsck shows a new one again19:27
← Jackneill left19:28
→ mateusz__ joined19:28
loops Thanatermesis: careful19:28
should prolly set a good HEAD before running that multiple times19:29
→ sq-one joined19:29
loops fsck will complain about objects that aren't referenced, even if they are intact19:29
_ikke_ depends on what message19:30
if it's dangling object, you can ignorei t19:30
loops yes19:30
_ikke_ ignore it19:30
Guest48858 → Cerise19:30
← Cerise left19:30
→ Cerise joined19:30
← mklappstuhl left19:30
Thanatermesis I should edit those files and put on it the last known object existing ? (before the last one) .git/refs/heads/master .git/logs/HEAD19:30
← johnw left19:31
→ mmlac joined19:31
← milind left19:32
→ johnw joined19:32
→ milind joined19:32
→ jcontra joined19:33
→ boombatower joined19:33
← AaronMT left19:33
← chamunks left19:33
loops the last line of .git/logs/HEAD should show you the sha1 you need...19:34
← rgrinberg1 left19:35
→ rgrinberg joined19:35
Thanatermesis loops, it actually shows me one of the (2 firsts) corrupted objects that i have removed19:35
loops which you can drop into .git/refs/heads/master19:35
← toky left19:35
Thanatermesis btw i have a backup that i can recover to the first state19:35
← pastime left19:36
← BSaboia left19:36
loops Thanatermesis: well if you can't use the last entry in the reflog, use the next-to-last19:36
Thanatermesis well, let me edit these two files to see what hapens19:36
loops Thanatermesis: well in truth, you don't need to edit the files to test19:36
you can just git log <sha1>19:36
to see if it will work19:36
← mixel left19:36
loops and if it does, then edit19:37
← johnw left19:37
→ johnw joined19:37
Thanatermesis yes, git log to the previous id works19:37
← mmlac left19:38
loops k, just drop that in .git/refs/heads/master19:38
and you're away to the races... rerun.. fsck19:38
Thanatermesis git fsck still shows a corrupted object, looks like its other one that was not the last-one19:39
i remove it from the .git/refs/heads/master list ?19:39
← taichunmin left19:39
loops Thanatermesis: not sure exactly what you're seeing. Maybe it's one of the objects you already deleted?19:40
Thanatermesis: you said that git log <sha1> works... does git log master now work?19:40
Thanatermesis loops, seems like no, each time i deleted an object it says that another (new) is corrupt19:41
yes, works :)19:41
"git log master"19:41
← mechanicalduck left19:41
loops Thanatermesis: okay.. do this.. git init a new repo, and git clone from this repaired one...19:41
→ guampa joined19:41
← KillYourTV left19:41
loops Thanatermesis: only the good objects will be copied...19:42
Thanatermesis mmh, let me see19:42
loops forgit the git init.. just git-clone obviously... i'm getting pulled away here.19:43
back in a bit19:43
← rudak left19:43
← jceb left19:43
→ rudak joined19:43
→ _iron joined19:43
→ mechanicalduck joined19:43
→ KillYourTV joined19:44
→ jceb joined19:44
[-jon-] is there a way to speed up git-clone over the network? limit history or such?19:45
← malkomalko left19:45
→ conan_ joined19:45
Thanatermesis http://www.enlightenment.org/ss/e-51b4db71ea5c17.13319255.jpg loops19:46
← ThomasLocke left19:46
→ nsanden joined19:46
← nbouscal left19:46
Thanatermesis mmh, resizing the terminal seems like included old text, tehre's no "git init" on this attempt19:46
canton7 [-jon-], there's --depth, but repos cloned with that option have lots of limitations19:47
[-jon-] its mostly network limited19:47
← guampa left19:48
loops Thanatermesis: you're only getting loose object messages? those aren't corrupt objects.. nothing to worry about.19:48
Thanatermesis loops, the screenshot is not very good made, but see the next line where says "fatal", it says a corrupt19:48
breaked on the next line19:49
← fobban left19:49
← ej left19:49
loops Thanatermesis: ah yeah, my bad...19:51
→ jhoepken joined19:51
loops Than, in the repaired repo... try git fsck master19:51
→ iveqy joined19:51
Thanatermesis same messages about the broken objects19:52
→ madewokherd joined19:52
loops Thanatermesis: k, that's not good news19:52
← a-l-e left19:53
→ jkubecki joined19:53
loops Thanatermesis: Appears some of those objects are empty files.. no data19:53
← Droolio left19:53
loops Thanatermesis: is git log master^ any better?19:54
Thanatermesis mmh, it shows the log but in the end it shows the error19:54
well, maybe that is my prompt, idk19:54
(zsh prompt that shows some git info)19:55
loops git fsck master^19:55
→ Cykey joined19:55
loops oh.. zsh... git fsck 'master^'19:55
Thanatermesis same error messages19:55
← jceb left19:55
loops yikes.. this wasn't just a few objects lost then i fear19:56
← aendrew left19:56
→ guns joined19:57
← x15 left19:57
loops Thanatermesis: you might want to go back to the original repo and try git log <sha1>19:58
← jkubecki left19:58
loops Thanatermesis: just to rule out a problem with your copy19:58
where sha1 is the one you recovered from the reflog19:58
but it's seeming like there was a lot more damage to your repo than a few uncommitted objects19:59
← nsanden left20:00
Thanatermesis well, i can just remove it and create a new one too20:00
new git repo from scratch20:01
loops probably diminishing returns on trying to save this one then..20:01
but it's a bit troubling... shouldn't lose a repo just because of a unplanned shutdown20:02
← marhaban1 left20:02
→ pantsman joined20:03
Thanatermesis loops, unreliable filesystems :)20:03
← jnavila left20:03
Thanatermesis im going to move back to my old and safe reiserfs because of that20:04
Thanatermesis of course, dont want to start a FS-flamewar20:04
loops which FS was this one?20:04
Thanatermesis ext420:04
maybe i should have not used the option nobarrier, not sure20:05
→ ncd|leeN joined20:06
← donnib left20:07
→ lindenle joined20:07
← leeN left20:09
→ hwrd|work joined20:12
← barneybook left20:14
← lindenle left20:14
→ breck7 joined20:14
→ barneybook joined20:14
← mechanicalduck left20:14
← danicampa901 left20:15
→ mechanicalduck joined20:15
→ raisen joined20:17
→ njbair joined20:17
← s00pcan left20:18
← hwrd|work left20:18
→ worblehat joined20:20
→ Bosox20051 joined20:21
→ chris64 joined20:23
→ frenchlaf joined20:24
← frenchlaf left20:24
→ mckendricks joined20:24
← lresende left20:24
← _iron left20:25
← paul424 left20:25
→ ej joined20:25
← rednaks left20:25
← Milossh left20:26
← chris64 left20:26
→ rednaks joined20:26
→ joebobjoe joined20:27
← asteros left20:29
← ej left20:29
→ hwrd|work joined20:30
→ _alex_kag_ joined20:32
← jhoepken left20:32
← mechanicalduck left20:32
→ Sajaki joined20:32
_alex_kag_ /msg gitinfo20:32
← daanavitch left20:33
→ ozette joined20:33
→ trusktr joined20:33
→ mechanicalduck joined20:33
joebobjoe I think it is sad that git does not fix its flaws because "that's what people are used to"20:35
→ aendrew joined20:35
→ thierryp joined20:36
← Guest67291 left20:36
iveqy joebobjoe: what are you thinking about?20:37
charon Thanatermesis: the description of the barrier options in 'man mount' leads me to believe that you lose all data consistency guarantees in the event of a crash if you use 'nobarrier', unless your device has fancy battery-backed caches20:37
Thanatermesis: that being said, there were also some recent ext4 bugs related to some of the more esoteric options20:38
(not sure if that included 'nobarrier', but it's possible)20:38
joebobjoe iveqy: Stuff involving imporvements to the CLI20:38
→ JaredMiller joined20:38
← Sosumi left20:38
← jcontra left20:38
← rednaks left20:39
→ sgambino joined20:39
→ Guest67291 joined20:40
→ s00pcan joined20:41
→ jcontra joined20:41
← thierryp left20:41
← sgambino left20:41
Thanatermesis charon, yeah, i have set this option in the past because im using a good SSD device (with batteries) after to read in some howtos somewhere, but im thinking now that it maybe means a different "battery" system, not sure20:41
bremner joebobjoe: well, the people who are unhappy don't often contribute patches afaict20:42
→ rednaks joined20:43
joebobjoe bremner: Ah okay20:43
→ barq joined20:44
← s00pcan left20:44
← howlymowly left20:45
→ s00pcan joined20:46
→ danicampa901 joined20:47
← swiety left20:47
← rudak left20:49
→ rudak joined20:49
← JaredMiller left20:49
← aendrew left20:49
→ kenperkins_ joined20:49
← kenperkins left20:50
→ mechanicalduck_ joined20:50
← irqq left20:53
← mechanicalduck left20:53
mechanicalduck_ → mechanicalduck20:53
→ jkubecki joined20:54
← raisen left20:54
→ flippo joined20:54
← jayfk left20:55
← barq left20:56
→ barq joined20:56
← noxd left20:56
← OliverJAsh left20:57
← apelisse left20:57
← darnir left20:58
← ellio left20:59
← scottrowe_ left20:59
← jkubecki left20:59
→ scottrowe_ joined20:59
← hinst left20:59
← s00pcan left21:00
← rednaks left21:00
→ s00pcan joined21:01
→ lindenle joined21:01
→ rednaks joined21:01
← s00pcan left21:01
→ s00pcan joined21:01
← chrob left21:02
← neutrino left21:02
← joebobjoe left21:03
← w0ng left21:03
← whowantstolivef1 left21:04
→ raisen joined21:06
→ FreezingCold joined21:06
→ joebobjoe joined21:07
← joebobjoe left21:07
→ LuxuryMode joined21:09
← hwrd|work left21:12
← AdrienBrault left21:14
→ chrob joined21:14
→ shentino joined21:17
← shentino left21:17
→ shentino joined21:17
→ ntus1017_ joined21:17
→ kapowaz joined21:17
← FreezingCold left21:17
→ zacts joined21:17
→ hwrd|work joined21:18
← hwrd|work left21:19
kapowaz hi there. I have a remote which I need to push to which apparently has changes I've not merged in my local master (from months ago). I don't care about any of these changes. How can I forcibly push my current master to this remote, even if it means overwriting stuff?21:19
i.e. “Updates were rejected because the tip of your current branch is behind its remote counterpart. Merge the remote changes (e.g. 'git pull') before pushing again.”21:19
cbreak kapowaz: beg who ever manages that remote for force push rights21:20
kapowaz: then force push21:20
kapowaz I own the remote too21:20
is that as simple as git push -f ?21:20
cbreak yes21:20
kapowaz tries that21:20
← ntus1017 left21:20
← gasbakid_ left21:21
← zacts left21:21
→ FreezingCold joined21:21
kapowaz no good. It's a heroku app, so I think there's some other bollocks it needs.21:22
kapowaz cries21:22
kapowaz okay, thanks anyway21:22
← system64 left21:22
→ guampa joined21:22
Guest67291 → DrBenway21:23
← nicedice left21:23
← LuxuryMode left21:23
→ nicedice joined21:24
← jgh left21:24
→ ej joined21:25
flo`_ → flo`21:26
thiago kapowaz: configure the remote server to allow remote pushes21:26
kapowaz it does.21:26
thiago kapowaz: do you have shell access to the remote repository?21:26
kapowaz nope21:26
thiago kapowaz: !repro the full output from the push -f21:26
gitinfo kapowaz: Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session -- or, even better for complex issues, design a minimal case in which your problem can be reproduced, and share it with us. This will help immensely with troubleshooting.21:26
kapowaz it's okay, I've decided to nuke it and sort it out by restoring the app21:26
→ zacts joined21:26
kapowaz since the only purpose for the remote is to deploy (it's a heroku app)21:26
thiago I don't know what a heroku app is21:27
kapowaz heroku.com21:27
it uses git for deployment21:27
i.e. git push will deploy your code21:27
thiago sounds like a bad idea21:27
git is not a deployment tool21:27
← FreezingCold left21:28
← sq-one left21:28
← ntus1017_ left21:28
kapowaz it works very well, provided you have good git discipline, which I didn't around the time I started using it for this app.21:28
← kfogel left21:28
→ ntus1017 joined21:29
← mechanicalduck left21:29
← danicampa901 left21:30
← ej left21:30
← mribeirodantas left21:30
→ noxd joined21:31
← barq left21:31
→ barq joined21:32
← mckendricks left21:32
→ LuxuryMode joined21:33
→ toky joined21:34
← phipes left21:36
← barq left21:37
→ barq joined21:37
← R1KRDO left21:37
→ rgb247 joined21:38
rgb247 how can I reverse a git stash?21:38
thiago reverse?21:39
osse rgb247: probably 'git stash pop' but I'm not quite sure what you mean21:39
thiago do you mean re-apply the changes that you stashed away?21:39
→ FreezingCold joined21:40
→ russellw joined21:42
→ Balliad joined21:43
← raisen left21:44
← uidzer0 left21:44
→ diegoviola joined21:45
→ BSaboia joined21:48
← Sajaki left21:48
← barq left21:49
← jaequery left21:52
→ jkubecki joined21:54
← ColoradoPods2 left21:54
← Bosox20051 left21:57
→ lresende joined21:57
→ braisarias joined21:57
← jkubecki left21:59
→ epta joined22:00
← DrBenway left22:00
→ nbouscal joined22:00
→ asteros joined22:01
← asteros left22:03
→ Mamoru_ joined22:03
→ amcsi_ joined22:03
← Mamoru left22:03
Mamoru_ → Mamoru22:03
→ mixel joined22:05
← noxd left22:05
← crustie left22:06
← amcsi left22:06
← fluxdude left22:06
← stamina left22:06
← banzoune1 left22:07
← s0x_ left22:07
← myme1 left22:07
→ s0x joined22:07
→ banzounet joined22:07
→ stamina joined22:07
→ myme1 joined22:07
→ asteros joined22:07
← Adlai left22:08
← rednaks left22:09
← drizzd left22:10
← pantsman left22:10
← Dougie187 left22:11
← asteros left22:14
← shruggar left22:15
→ itatitat joined22:17
→ system64 joined22:17
← chrob left22:18
→ bigkm joined22:18
← Cykey left22:19
→ asteros joined22:20
← worblehat left22:20
← doener left22:20
← SeptDePique left22:26
→ ej joined22:26
← Tronic left22:28
→ aendrew joined22:29
→ dsdeiz joined22:30
← ncd|leeN left22:30
→ Tronic joined22:31
← ej left22:31
← Nicholas_Rishel left22:33
← zz_pinage404 left22:34
← itatitat left22:35
← aendrew left22:41
← Davey left22:41
[-jon-] I'm a bit confused -- how do people handle an IDE and git? it seems like switching branches really would wreak havoc on your IDE because modules appear and disappear, and compiled code is likely wrong, etc22:41
→ mklappstuhl joined22:41
← aard_ left22:43
→ salimfadhley joined22:43
→ noxd joined22:43
osse [-jon-]: why does the compiled code matter?22:43
[-jon-] it'll be out of sync with the actual code, osse22:44
youll have to recompile or be in a funky state22:44
salimfadhley Question: I've been working on something, committing locally (but not pushing) - is there a way to push all my changes to a new branch?22:44
→ aendrew joined22:44
osse [-jon-]: th22:45
[-jon-] I'm mainly concerned about modules, etc22:45
osse [-jon-]: it's true that it will be out of sync. i just don't see how it's a problem.22:45
[-jon-] it seems like I would WANT untracked files to stay with branches, just not be committed..?22:45
osse: it makes switching of branches significantly more heavyweight22:46
→ Davey joined22:46
→ w0ng joined22:46
osse are the two branches very different?22:46
[-jon-] and if you dont clean before switching, youll have compiled code which the actual code is no longer in your branch22:46
not terribly difficult22:47
cmn this is no different from removing or changing a file w/o switching branches22:48
osse when you switch branches git only touches the files it needs to, so you only need to recompile the changed files22:48
cmn if a compiled object is out of date, the build system should rebuild it22:48
osse no need to clean22:48
russellw Visual C++ handles it fine, it automatically notices changes22:48
[-jon-] unless thats a new module22:48
← Mystus left22:48
[-jon-] in that case its not even part of your build plan anymore22:48
russellw And from what I remember of when I was using Netbeans, I don't think it had any difficulty either22:48
[-jon-] it just seems awkward22:48
cmn but that's not an issue that comes from git22:48
[-jon-] its an issue that comes from using git22:49
cmn if your build system can't handle that, then have your IDE do it22:49
osse [-jon-]: usually the building stuff is versiond alongside the code itself, so thst shouldn't be a problem22:49
cmn but this is exacty the same as if you remove a file, no different when using it22:49
← Balliad left22:50
[-jon-] it seems like switching branches will take like 5 minutes22:50
having to fully build, reindex, etc22:50
← ntus1017 left22:50
cmn then don't22:50
move to a different directory if it's so unconfortable22:50
osse if the build system gets confused from stray object files i'd say that's a flaw in the build system22:51
[-jon-] osse: its not that it gets confused22:51
its that youll just have a ton of shit hanging around22:51
← Tronic left22:51
osse you shouldn't need to rebuild everything.22:51
← russellw left22:51
← lresende left22:51
osse does that matter?22:51
← FreezingCold left22:51
[-jon-] it just seems like there should be a way to have untracked files stay with the branch22:52
cmn that goes against what a branch ish22:52
if you need that, then switch directories22:52
[-jon-] when I checkout a new branch, I would at least expect to be at that branch's state, not some weird state with old untracked files22:52
→ Cykey joined22:53
← theshadow left22:53
[-jon-] there's no way to just locally track files, but not push them, is there? (at least a non-awkward way)22:53
EugeneKay [-jon-] - "no"22:53
osse git cannot know the proper state of untracked files. for that you would need to track them22:53
EugeneKay You can maintain a Public and a Private branch, and cherry-pick stuff from Private over to Public, then only pus hthat.22:54
[-jon-] seems like more trouble than what its worth22:54
osse [-jon-]: maybe you want to follow each git checkout with a cit clean?22:54
but then you *definitely* need to rebuild everyhing22:54
[-jon-] heh, and hope I didn't forget to track something important22:55
→ jkubecki joined22:55
milki if you arent tracking it, its not important22:55
!22:55
→ OliverJAsh joined22:55
[-jon-] I was hoping to use git-svn for feature branches, but it just seems like a lot of headache22:55
osse even if git could magically keep an d22:56
adhawkins-away → adhawkins22:56
← ferdev left22:56
osse even if git could magically keep an eye on untracked files the modification times would still change so you would still need to rebuild22:56
→ ferdev joined22:56
osse heh, and i still don't see the meed for rebuilding in the first place22:57
[-jon-] modification times dont matter with maven22:57
→ ellio joined22:57
osse oh22:58
cmn what do you mean they don't matter?22:58
→ babilen_ joined22:58
[-jon-] maven is generally poor about detecting changes in other modules and determing what to rebuild22:58
osse maybe it uses checksums instead. some build systems do that e.g. scons22:58
[-jon-] doing a clean install isn't THAT slow, just somewhat inconvenient22:59
I'm more concerned about IDE-specific files and modules23:00
→ Joshi3s` joined23:00
← jkubecki left23:00
← Jahm left23:00
← babilen left23:00
osse if you are concerned i suppose there is *aomething* that gets confused about this messy state23:00
← noxd left23:01
← lindenle left23:01
[-jon-] heh, doing a clean would probably screw up an IDE23:01
← mklappstuhl left23:02
adhawkins → adhawkins-away23:02
← rudak left23:03
osse what doesn't? :p23:03
EugeneKay SATA is better23:03
← asteros left23:06
← gittish left23:07
← stamina left23:09
→ lindenle joined23:11
→ cagedwisdom joined23:12
osse EugeneKay: "I'm gonna make him a pull request he can't refuse"23:13
→ mckendricks joined23:14
milki gah, /me kicks git smart protocol23:16
linux is sending me a RST packet instead of trying to FIN, ACK -.-23:16
← thiago left23:17
→ mmlac joined23:17
→ rlubke joined23:18
→ noxd joined23:19
→ devoldmx joined23:20
cmn that wouldn't be the protocols fault, most likely23:21
← rlubke left23:21
→ SexyGirlG joined23:21
cmn I've seen that when you hit the remote end's connection time limit23:21
← nbouscal left23:22
← amcsi_ left23:23
→ Blahhh joined23:25
→ natechan joined23:26
← Blahhh left23:26
← Ingmar left23:26
→ ej joined23:26
→ Blahhh joined23:27
← Blahhh left23:27
← penk left23:27
← LuxuryMode left23:27
→ Etheco- joined23:27
→ Blahhh joined23:28
← SirCmpwn left23:28
→ OMGOMG_ joined23:28
← Blahhh left23:28
→ relipse joined23:29
← Etheco left23:30
→ LuxuryMode joined23:30
milki its localhost though23:30
and its tcp23:30
→ Jahm joined23:30
milki well, i need to post these caps and have someone else look at them -.-23:31
→ SirCmpwn joined23:31
milki hm23:31
i dont know where to ask >.<23:31
← ej left23:31
→ wbill joined23:32
unreal_ → unreal23:32
cmn hm, interesting23:32
← wbill left23:32
cmn did git-daemon complain?23:32
← Goplat left23:32
milki hmm, im not sure actually23:33
i wonder where the code is calling git-daemon...23:33
← DonRichie left23:33
→ wbill joined23:34
← wbill left23:34
cmn this is in some app wrapping git-daemon or do you mean client-side?23:35
→ wbill joined23:35
→ R1KRDO joined23:35
milki this is dulwich23:35
← system64 left23:36
milki so the client side is definitely pure-python23:36
im not sure what the "server" side is during the test23:36
← wbill left23:36
milki posted my traces here: https://bugs.launchpad.net/dulwich/+bug/106308723:36
i know the temp repo is created with git init23:36
milki looks for the server side invocation23:36
← aendrew left23:37
milki it might be calling git-receive-pack directly23:37
→ wbill joined23:37
→ Dougie187 joined23:37
← wbill left23:37
← R1KRDO left23:38
cmn hm, wouldn't dulwhich test against its own server?23:38
→ wbill joined23:38
milki this is the compatibility tests23:38
dean → dean|away23:38
milki so it should be testing against git23:38
→ aendrew joined23:39
milki ah yes, it pulls up a git-daemon23:39
→ phipes joined23:39
← ozette left23:39
milki i should try latching on to its stderr then23:40
cmn run it with --verbose, then it will tell you (more or less) why it's killing the connection23:40
← sajan left23:40
milki its already run with --verbose23:41
but probably the test framework is hiding some of the output23:41
hm...maybe i manually run the git-daemon instead23:43
cmn that run_git_or_fail doesn't sound like a chatty function23:44
the way I'm reading the test code, it's going to skip it if it's already running23:44
in https://github.com/milki/dulwich/blob/send_pack/dulwich/tests/compat/test_client.py#L22623:45
← donmichelangelo left23:45
→ donmichelangelo joined23:46
→ Fandekasp joined23:47
← kukks left23:47
← FalsAlarm left23:48
→ FalsAlarm joined23:49
← nafg left23:49
→ kukks joined23:49
← kukks left23:49
→ kukks joined23:49
← FalsAlarm left23:49
← Chillance left23:51
→ Internet13 joined23:52
→ R1KRDO joined23:52
← R1KRDO left23:52
[-jon-] if you use --depth for a clone, why can't you pull/push it o_O23:54
→ jkubecki joined23:56
← dsdeiz left23:56
Dougie187 [-jon-]: http://stackoverflow.com/questions/6941889/is-git-clone-depth-1-shallow-clone-more-useful-than-it-makes-out23:56
specifically --- http://stackoverflow.com/questions/6900103/why-cant-i-push-from-a-shallow-clone23:56
← SirCmpwn left23:57
→ SirCmpwn joined23:58
→ Tronic joined23:58
← freestyl3r left23:58
→ CodeShark joined23:58
→ freestyl3r joined23:59
[-jon-] Dougie187: so it works if you have sufficient history?23:59

Logs Search ←Prev date Next date→ Channels Documentation