Time Nick Message 06:09 ventz git has officially pissed me off 06:09 ventz pretty much unless you are using github, it's a piece of shit software when it comes to server compatability 13:27 pdurbin ventz: server incompatibility with git? 13:43 pdurbin agoddard: https://github.com/agoddard/PassNG seems interesting but. . . lots of dependencies 13:57 Pax morning 14:05 pdurbin morning 14:30 pdurbin finally getting around to adding "--type-set=puppet=.pp" to ~/.ackrc 14:32 shuff \o/ 14:34 pdurbin shuff: much stuff in your .ackrc? 14:35 shuff $ cat .ackrc 14:35 shuff --type-set=puppet=.pp 14:35 shuff --type-set=rpm=.rpm,.spec 14:35 shuff --type-set=spec=.spec 14:35 shuff not too much 14:35 pdurbin yeah, it's pretty usable out of the box 14:40 pdurbin shuff: what hint? (from yesterday) 14:57 shuff the hint is "yes, you should" :) 14:58 pdurbin ah. ok :) 15:22 pdurbin hmm. nice article on overcommitment of memory (and cpu) with kvm http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization_Administration_Guide/sect-Virtualization-Tips_and_tricks-Overcommitting_with_KVM.html 16:39 ventz pdurbin: yea, it reminds me of SVN 10+ years ago 16:39 ventz so basically, to me only 2 methods of "exposure" are realistic - http and ssh 16:40 ventz in order to have http -- easy enough, drop in a webroot -- ok, strange, why not have it like an app (like svn) -- oh wait, that's where gitweb comes in -- hack on top of a poorly made decision, ok, so now how do we write back? yet another hack, webdav, ok, sort of works, most of the time, it's super inneficient, and oh wait, git has problems on it's own too that cause other problems (put bugs and pack -- which uncover an apache bug...story for another 16:41 ventz so now, ssh -- ok, works PERFECTLY, but doesn't play well with anything else? why? because the other methods are hacks on top of poorly made decisions 16:41 ventz so, in order for me to: 16:41 ventz 1.) provide ssh access for making changes, 2.) provide http access (which stays updated after ssh pushes) for get access, and 3.) provide a web view -- something always ends up breaking 17:32 pdurbin ventz: why do your users need http access to git? 17:48 pdurbin just do it like github does. that's what i'm trying to do here. and it's working, more or less 18:17 pdurbin hmm. How do large organizations with large turnover and many systems perform account provisioning? - Server Fault - http://serverfault.com/questions/198389/how-do-large-organizations-with-large-turnover-and-many-systems-perform-account 18:56 ventz pdurbin: i have a box that i don't want to give users ssh accounts to (mostly b/c it's ldap based, has other access, complicated, etc... -- htaccess file is just much easier) 18:57 ventz The problem comes when you push with ssh, it breaks gitweb, and if you fix the permissions there, it breaks git over http -- it's outdated 18:57 ventz you do the git update-server-info, now it breaks the other one again... 18:57 ventz i have the post hook, still nothing 18:58 pdurbin is there a howto you're following to set up the http access? i've never done this... 18:58 ventz i used stuff on the web (set this up > 1 year ago), but last night I found this: http://www.jedi.be/blog/2009/05/06/8-ways-to-share-your-git-repository/#apachehttp 18:58 ventz which seems pretty good 18:59 pdurbin yeah, i've seen this blog post before 19:05 pdurbin ventz: this is getting a bit dated, but you might be interested in it: http://code.google.com/p/support/wiki/DVCSAnalysis 19:06 pdurbin it explains why google chose to support hg rather than git, in part due to better http support in hg 19:06 pdurbin support hg projects on code.google.com that is 19:27 ventz pdurbin: thanks. that was interesting 19:28 ventz i think i am going to move my git repo to svn sadly 19:28 ventz i'ts just too frustrating 19:44 pdurbin so sad