<prahal>
indirect: thanks . I saw your comment closing the bug report I opened about above issue. I replied (namely explaining that I was told on #ruby that one cannot rely on array .hash to prevent collision
<indirect>
prahal: by whom?
<indirect>
oh I see
<indirect>
I have no idea who havenwood is, but that is incredibly sucky if true
<prahal>
havenwood
<prahal>
agreed, so I should resume my reading the rb_hash code I guess ,)
<indirect>
prahal: afaik, there is a _very_ strong expectation that there will not be collisions
<indirect>
because iirc that .hash value is used for hashes
<indirect>
and if you have collisions the entire Hash data structure stops working
<havenwood>
apeiros: hash values are *bound* to collide.
<indirect>
havenwood: who are you talking to?
<havenwood>
oh, just quoting the irc log
<indirect>
oh
<havenwood>
sry, that was unclear >.>
<indirect>
yes
<indirect>
mathematically
<indirect>
there are less hash values than there are objects
<indirect>
the goal is to make it extremely unlikely that they will collide
<indirect>
and in fact, ruby on any other platform does not collide
<havenwood>
i just wasn't sure, makes sense
<havenwood>
upstream problem
<indirect>
yes
<indirect>
I don't think that the Hash class will function correctly
<indirect>
so this is not actually patchable in bundler
fromonesrc has joined #bundler
<indirect>
prahal: havenwood: changed my mind after talking with some other peopl
<prahal>
replied , no success . I will turn to the ruby tracker. Thanks . I am in a better position to carry this issue to completion .
axl_ has quit [Quit: axl_]
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
fromonesrc has quit [Ping timeout: 272 seconds]
fromonesrc has joined #bundler
patcon has quit [Ping timeout: 272 seconds]
patcon has joined #bundler
fromonesrc has quit [Ping timeout: 245 seconds]
woollyams has quit [Ping timeout: 252 seconds]
woollyams has joined #bundler
lolmaus has joined #bundler
<lolmaus>
Hi! My `bundle install` routine freezes at "Installing libv8 (3.16.14.3)". Can you please help me with this?
woollyams has quit [Ping timeout: 252 seconds]
woollyams has joined #bundler
JordanFaust_ has joined #bundler
<lolmaus>
Okay, i should've been more patient
<indirect>
lolmaus: lol
<lolmaus>
indirect, it took more than 30 minutes without any indication
<lolmaus>
Just for that single item
<indirect>
what do you think is should to indicate?
<indirect>
it's running configure and make in the background
<indirect>
I have no control over what you decide to install :)
<lolmaus>
indirect, is there a verbose mode?
<indirect>
lolmaus: sure. you activate it with --verbose
<indirect>
it's not going to tell you much more though
<indirect>
it's just compiling :\
<indirect>
there is no case where it prints out the entire build output
<lolmaus>
indirect, when make runs manually it produces a lot of output and you can tell when it stalls.
<lolmaus>
Okay maybe i should just trust it more.
<lolmaus>
It seems to be way more reliable than `npm install` anyway.
JordanFaust_ has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
x1337807x has joined #bundler
lolmaus has quit [Read error: Connection reset by peer]
lolmaus has joined #bundler
patcon has quit [Ping timeout: 248 seconds]
rjhunter has quit [Remote host closed the connection]
rjhunter has joined #bundler
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
retro|cz has quit [Ping timeout: 240 seconds]
woollyams has quit [Ping timeout: 252 seconds]
samphippen has joined #bundler
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
samphippen has joined #bundler
samphippen has quit [Client Quit]
Paresh has joined #bundler
samphippen has joined #bundler
mfoo has joined #bundler
Paresh has quit [Ping timeout: 250 seconds]
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
JordanFaust_ has joined #bundler
prahal has quit [Ping timeout: 272 seconds]
samphippen has joined #bundler
prahal has joined #bundler
JordanFaust_ has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
JordanFaust_ has joined #bundler
shtirlic has joined #bundler
JordanFaust_ has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
vanderkerkoff has joined #bundler
<vanderkerkoff>
Hello everyone
<vanderkerkoff>
I'm having alot of fun trying to get bundler to work during a capistrano deploy
<vanderkerkoff>
The server we're deploying to is behind a proxy
<vanderkerkoff>
When I'm on the server, I can run bundle install from within a folder with the exact same Gemfile in it, no problem
<vanderkerkoff>
but in the deploy it keeps failing Could not fetch specs from https://rubygems.org/
fromonesrc has joined #bundler
samphippen has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
rjhunter has quit [Remote host closed the connection]
<vanderkerkoff>
someone throw me a bone!!
<mfoo>
never used cap, sorry
<mfoo>
if it works with 'bundle install' from the terminal, it's probably an environment variable or capistrano problem
<dwradcliffe>
are you using bundler/capistrano?
<mfoo>
if it relies on the http_proxy variable, you should check if it's set in your current shell where it succeeds, and in the cap stage
<mfoo>
probably cap starts a non login shell and doesn't source whatever file sets $HTTP_PROXY
<vanderkerkoff>
Yeah, we're using bundler/capistrano
<vanderkerkoff>
you got me thinking
<vanderkerkoff>
maybe I should set it in the capistrano deploy script
<mfoo>
well, where do you set HTTP_PROXY when you run it from the shell? ~/.bashrc?
<vanderkerkoff>
in etc/profile
<vanderkerkoff>
the proxy is setup fine, when I'm logged in
<vanderkerkoff>
like you say, it looks like capistrano is not sorucing the http-proxy value
<mfoo>
/etc/profile is not invoked for non login shells
<mfoo>
you need to start bash -l
<vanderkerkoff>
I'll try adding it to .bashrc
<vanderkerkoff>
this is not one of my servers, btw
<vanderkerkoff>
I expect everyone says that when something turns out to be shit :-)
<dwradcliffe>
haha
<mfoo>
that won't be sourced in non login shells either
<mfoo>
err
<vanderkerkoff>
ok, so, what will be sourced in non login shells?
<mfoo>
I meant the opposite
<vanderkerkoff>
I think I'm going to have to find a way to set it in capistrano
<mfoo>
~/.bashrc should be used in non login shells
<mfoo>
~/.bash_profile is for login shells
<vanderkerkoff>
ahh
<vanderkerkoff>
ok
<vanderkerkoff>
I'll give it a whirl
<mfoo>
but it might be easier to set in capistrano
<vanderkerkoff>
yes, you are right, I wont have to rely on the server team getting it right every time then
<mfoo>
see the "Invocation" section in `man bash` for more info
<vanderkerkoff>
:q!
<vanderkerkoff>
woops
<vanderkerkoff>
that could have been worse :-)
<mfoo>
no saved changes here :)
<mfoo>
*unsaved
<vanderkerkoff>
moving the http_proxy definitions into .bashrc worked
<vanderkerkoff>
thanking you greatly
<mfoo>
do you have puppet/chef/something to manage that change?
<mfoo>
just consider when deploying a new application to a new server you'll have to remember this :)
<mfoo>
one reason why doing it inside capistrano would be better, unless the proxy is different per site!
<mfoo>
good that it works now
<vanderkerkoff>
I'll look into capistrano once I've told the people screaming at me for this site that its up
chouhoulis has joined #bundler
chouhoul_ has joined #bundler
chouhoulis has quit [Read error: Connection reset by peer]
vanderkerkoff has quit [Quit: vanderkerkoff]
axl_ has joined #bundler
<mfoo>
I would like to use a different gem server in production than development. Preferrably with the same Gem file. Can I change the "source" command based on an if statement inside the Gemfile?
<dwradcliffe>
I *think* the source is part of the lock file, so that won't work
<mfoo>
I am wondering if the rails environment is in scope at that point, and whether the Gemfile is interpreted or just parsed
<dwradcliffe>
Gemfile is way before rails is loaded
axl_ has quit [Quit: axl_]
<mfoo>
actually yes, that's obvious
<mfoo>
rails isn't loaded as part of bundle install, how could it know
<dwradcliffe>
right
<mfoo>
I don't really want to put my gemfile in puppet, as it changes a lot
<mfoo>
I could modify it with a script before running bundle install I guess
<dwradcliffe>
why do you need different sources?
<mfoo>
production servers have no access to the internet, we have our own gem server for production deployments
<dwradcliffe>
oh that's right, we were talking about this the other day
<mfoo>
yep that's me
<dwradcliffe>
I really don't think that will work
<dwradcliffe>
I'm assuming you can't access the production gem server in dev?
<mfoo>
indeed
<dwradcliffe>
you could try just using and ENV variable
<mfoo>
I'll have to sed the Gemfile and Gemfile.lock, there seems to be no command line parameter for it (and why should there be, it's in the gemfile)
<mfoo>
env variable would work too, if it's evaluated (does standard ruby work)
<dwradcliffe>
yes
<mfoo>
i've previously had an annoying issue with ruby evaluation insdie the gemfiel
<dwradcliffe>
we do it all the time
<mfoo>
there's an issue with exception handling, I think
<dwradcliffe>
although it's usually a sign that you're doing something funky :)
<mfoo>
I can have it default to the dev environment, so people won't have to make changes
<mfoo>
this is a better solution than sed :)
<mfoo>
thanks
<dwradcliffe>
np
<dwradcliffe>
I still think you may have a problem with the lock file
<mfoo>
this also includes the remote source
<mfoo>
as for the ruby code in gemfile thing, I had a version of ruby with no readline support enabled, so I had the following in my Gemfile: begin, require 'readline' rescue gem 'rb-readline' end, with the intention of requiring the rb-readline gem if the user didn't have the right version of ruby
<mfoo>
my rescue block never got executed, it bubbled up to bundler and it handled it
<mfoo>
which led me to believe some odd things happen in the gemfile, perhaps it's evaluated line by line
<mfoo>
(ignoring the fact that this would make Gemfile.lock different for people with different environments, causing issues if you're working in a group)
<mfoo>
that's not bundler's problem
patcon has joined #bundler
<dwradcliffe>
you will probably see "Cannot write a changed lockfile while frozen." but I'm not sure if it will work anyway
<mfoo>
I haven't set --frozen, all seems to be good
<dwradcliffe>
when you run on the server with --deployment it will be
<mfoo>
heh, maybe it would be easier to add an /etc/hosts entry for rubygems.org to my gem repo :D
<mfoo>
I will not do this.
greendot_ has joined #bundler
<greendot_>
Hello everyone :) . I'm trying to deploy my Sinatra app to a vagrant vm. I am also using Nginx and Phusion Passenger. Passenger is complaining about not being able to find the gems, even though I have checked in the entire vendor bundle. I am running Bundler.require in config.ru. Am I missing a step here?
ixti has joined #bundler
<havenwood>
greendot_: Assuming you're using the Nginx module Passenger, after you build Passenger it provides an Nginx .conf snippet to point the module at the Ruby you built it against. Has Passenger been built with your current Ruby? (If not `passenger-install-nginx-module` and follow instructions.) Or is it something else?
<greendot_>
We couldn't get gem install passenger to work so we've installed it as a package via apt-get
<greendot_>
Passenger seems to be working fine, we get the same error when we run ruby config.ru on the server
<greendot_>
server = vagrant
<greendot_>
Sorry about not clarifying that in my initial question
greendot_ has quit [Ping timeout: 250 seconds]
patcon has quit [Remote host closed the connection]
patcon has joined #bundler
patcon has quit [Ping timeout: 272 seconds]
axl_ has joined #bundler
patcon has joined #bundler
patcon has quit [Ping timeout: 246 seconds]
patcon has joined #bundler
<indirect>
mfoo: dwradcliffe: the beta version of bundler (1.5) includes a bundle config mirror setting that will do exactly what you want
<indirect>
without editing the gemfile
<dwradcliffe>
indirect: very cool
patcon has quit [Remote host closed the connection]
patcon has joined #bundler
patcon_ has joined #bundler
patcon has quit [Read error: Connection reset by peer]
fromonesrc has quit [Ping timeout: 265 seconds]
<mfoo>
indirect, excellent, thanks
mfoo has quit [Read error: Operation timed out]
dwradcliffe has quit [Quit: Bye]
dwradcliffe has joined #bundler
dwradcliffe has quit [Quit: Bye]
dwradcliffe has joined #bundler
fromonesrc_ has joined #bundler
fromonesrc_ has quit [Ping timeout: 245 seconds]
patcon has joined #bundler
patcon_ has quit [Read error: Connection reset by peer]
patcon has quit [Remote host closed the connection]