[chef-dev] Re: 4/4 Opscode Code Review

[Joshua Timberman < >]

We ran long in the meeting, so we stopped COOK review short. Here is
some additional COOK tickets for your pleasure!

* To Merge:
- COOK-1114: correct File namespace in java::openjdk
- COOK-477: module compilation support for nginx::source
- COOK-873: better control via attributes for varnish

Note that COOK-477 introduces an attribute change for
node['nginx']['url'] -> node['nginx']['source']['url']. We thought
this would be more clear as to the intention. That ticket's pull
request is actually already merged, and the other two will be merged
soon too.

COOK-873 also resolves COOK-1091.

On Wed, Apr 4, 2012 at 3:02 PM, Bryan McLellan 
<
 >
 wrote:
> We realized today while looking at CHEF-2958 that some package
> providers take a filename in the source attribute and sometimes in the
> name attribute and that is unclear, and unfortunate. There isn't a
> whole lot to do about it right now without making a design change on
> the next major release.
>
> We've merged all the outstanding tickets for Chef. I'll go through
> todays results and ship a beta release of 0.10.10 before the end of
> the week so we can start testing it.
>
> To merge:
>  CHEF-2958 - gem_package install fails when gem_binary is specified
> and source is a local file
>  CHEF-3003 - Route provider does not create route files on redhat and fedora
>  CHEF-3008 - Route provider does not save all defined routes in
> redhat route configs
>  COOK-733 - add /etc/aliases template to postfix cookbook
>  COOK-856 - Add the vhost_combined LogFormat
>
> Other:
>  CHEF-2996 - Add ability to use a regexp for value_for_platform
> platform version specifier
>  Closed - Wontfix in favor of CHEF-2410
>  CHEF-1251 - FreeBSD package provider should not need the port tree
> to handle binary packages
>  Reopened - Avoid using rescue to blindly capture a failing command.
>  CHEF-3012 - Windows group provider is not idempotent
>  Reopened - Can we check for a group membership before we try to add
> a user to one?
>  COOK-743 - dynect::a_record has a typo // Dynect API expects
> "ARecord", not "A"
>  Reopened - Waiting until upstream library stabilizes
>  COOK-1148 - DMG Package provider should support the type "pkg"
>  Closed - Duplicate COOK-852
>  COOK-852 - dmg cookbook can target type "pkg"
>  Reopened - Use shell_out, combine contributions.
>  COOK-967 - passenger_apache2 relies on apache recipe to create a
> .load file but the apache2 module does not create one in debian based
> distros
>  Closed - Wontfix
>
> Andrea Campi is working on the refactor of the application cookbook,
> so I've assigned these to him:
>  COOK-657 - Ability to decouple application environment from 
> chef_environment
>  COOK-628 - Applications with recipes that have already been used
> silently fail.
>  COOK-625 - Allow other cookbooks to define application recipes
>  COOK-602 - allow _default environment default to be configurable
>  COOK-963 - application w/rails recipe fails for older rails 3.0.x
> apps. Should use 'bundle exec unicorn' to launch unicorn when app uses
> bundler
>  COOK-768 - application::rails passes the wrong environment to
> "bundle install --without"
>  COOK-786 - [PATCH] Application Cookbook Django recipe uses
> chef_environment as attribute rather than method
>
> --
> Bryan McLellan | opscode | technical program manager, open source
> (c) 206.607.7108 | (t) @btmspox | (b) http://blog.loftninjas.org



-- 
Opscode, Inc
Joshua Timberman, Technical Program Manager
IRC, Skype, Twitter, Github: jtimberman