Updates for Vagrant 1.3

[fgrehm]

From the changelog:

config.ssh.max_tries is gone. Instead of maximum tries, Vagrant now uses a simple overall timeout value config.vm.boot_timeout to wait for the machine to boot up.

That means we need to fix https://github.com/fgrehm/vagrant-lxc/blob/master/lib/vagrant-lxc/action/boot.rb#L27

Besides that we need to support the new ProvisionerCleanup action from hashicorp/vagrant@d4c7e20

[stucki]

I just upgraded Vagrant to 1.3.0 and got hit by this problem. For now, here's a hotfix solution:

diff --git a/lib/vagrant-lxc/action/boot.rb b/lib/vagrant-lxc/action/boot.rb
index 9edb163..b64e43b 100644
--- a/lib/vagrant-lxc/action/boot.rb
+++ b/lib/vagrant-lxc/action/boot.rb
@@ -24,7 +24,7 @@ module Vagrant
         def wait_for_boot
           @env[:ui].info I18n.t("vagrant_lxc.messages.waiting_for_start")

-          @env[:machine].config.ssh.max_tries.to_i.times do |i|
+          1.times do |i|
             if @env[:machine].communicate.ready?
               @env[:ui].info I18n.t("vagrant_lxc.messages.container_ready")
               return true

I'm wondering after all why these SSH connection attempts should fail once but work on the next try. In my opinion, it will either work, or fail forever...

[fgrehm]

@stucki tks for the hotfix :) as stated on a comment above that method, it was shamelessly stolen from the old version of VagrantPlugins::ProviderVirtualBox::Action::Boot and 1.3+ ships with a more robust built in action to deal with that stuff now.

I'll try to update vagrant-lxc to work on 1.3 later on tonight. I'm not sure how much effort it will take but I hope it won't be that hard, if it takes too much I might just drop support for vagrant < 1.3 on 0.6+ and keep the 0.5.x of the plugin series for 1.1 and 1.2 compatibility

[coveralls]

Coverage increased (+0.54%) when pulling e4ee20c on 136-updates-for-vagrant-1.3 into 7f3f7b1 on master.

[fgrehm]

Fixed on master! 🎆