Puppet Dashboard 1.2.13 is a maintenance and bugfix release of Puppet Dashboard.
This release is available for download at:
https://downloads.puppetlabs.com/dashboard/puppet-dashboard-1.2.13.tar.gz
Debian packages are available at
https://apt.puppetlabs.com
RPM packages are available at
https://yum.puppetlabs.com
See the Verifying Puppet Download section at:
http://projects.puppetlabs.com/projects/puppet/wiki/Downloading_Puppet
Please report feedback via the Puppet Labs Redmine site, using an
affected version of 1.2.13:
http://projects.puppetlabs.com/projects/dashboard
Documentation is available at:
http://docs.puppetlabs.com/dashboard/index.html
===============================
## Puppet Dashboard 1.2.13 Bug Fixes ##
===============================
(#17161) Fix stylesheet URLs for IE colorbox images
The colorbox CSS looks for the IE images in
/images/internet_explorer. The correct location is
/stylesheets/images/internet_explorer.
Update debian and redhat init script to use INT instead of term to stop
Previously the debian init script didn't pass --signal to the
start-stop-daemon, which defaults to TERM. Dashboard runs out of webrick by
default, which catches and ignores TERM, but will respond to INT and stop
sanely. Currently, running 'service puppet-dashboard stop' returns
[2012-03-27 10:24:17] ERROR SignalException: SIGTERM
/usr/lib/ruby/1.8/webrick/server.rb:91:in `select'
and then waits for the retry timeout, at which point
start-stop-daemon SIGKILLs
the ruby process. If --signal is set to INT in the init script, running
'service puppet-dashboard stop' returns
[2012-03-27 10:28:35] INFO going to shutdown ...
[2012-03-27 10:28:35] INFO WEBrick::HTTPServer#start done.
Exiting
which is the expected behavior.
The redhat init script suffered from a similar problem, adding -INT to the
killproc call speeds up the `service puppet-dashboard stop` call as killproc
doesn't need to wait and see if the process has ended, as INT is
responded to
immediately by webrick.
(#9296) Fix loading rack for rubygems >= 1.8
Rubygems version 1.8 broke some of the ways vendored gems used to
work in Rails
2. This commit updates the config/environment.rb to add all of the
gem libs to
the rails autoload_paths. It also updates the actionpack action_controller
library to not call out to gem explicitly. We know that the correct rack
version is being supplied in the vendored gems, so the following
require 'rack'
line will succeed.
This also munges the load path in config/boot.rb because boot.rb
doesn't load
environment.rb right away, and so can't find rack, due to the same
gem problem.
So this uses the same loop as in environment.rb, but instead of the rails
autoload_paths, it adds the libs to $:, or $LOAD_PATH.
===============================
## Puppet Dashboard 1.2.13 Changelog ##
===============================
Branan Purvine-Riley (1):
c2e4a4a (#17161) Fix stylesheet URLs for IE colorbox images
Matthaus Owens (3):
a3f2588 (#9296) Fix loading rack for rubygems >= 1.8
5378b46 Update debian and redhat init script to use INT instead
of term to stop
7a38e19 Updating CHANGELOG, VERSION for 1.2.13. Updating .gitignore
Moses Mendoza (1):
ee6d3c3 Update VERSION, CHANGELOG for 1.2.13-rc1