Bkr 1115 Allow options-file to be configured via ENV

[barrymw]

In order to run beaker acceptance tests unmodified behind a proxy, the host_env configuration needs to be externally settable (i.e. outside of the tests). As this is passed in to beaker via the --options-file, it makes (made?) sense to use expose this runtime parameter.

When used in conjunction with BKR-1114 pull request, it is possible to do the following:

git clone https://github.com/puppetlabs/puppetlabs-inifile.git
cd puppetlabs-inifile
export BEAKER_options_file=~/beaker-proxy.cfg
export DOCKER_BUILDARGS="HTTP_PROXY=http://10.0.2.2:3128/ HTTPS_PROXY=http://10.0.2.2:3128/     http_proxy=http://10.0.2.2:3128/       https_proxy=http://10.0.2.2:3128/"
PUPPET_INSTALL_TYPE=agent BEAKER_debug=true BEAKER_set=docker/centos-7   bundle exec rake -v beaker

Where ~/beaker-config.cfg contains:

{
  :package_proxy => 'http://10.0.2.2:3128/', 
  :host_env => {
    "http_proxy" => "http://10.0.2.2:3128/",
    "https_proxy" => "http://10.0.2.2:3128/",
    "HTTP_PROXY" => "http://10.0.2.2:3128/",
    "HTTPS_PROXY" => "http://10.0.2.2:3128/"
  }
}

[puppetlabs-jenkins]

Can one of the admins verify this patch?

[kevpl]

@barrymw apologies for the huge delay looking at this. Looks good to me, I'll put it through testing, & hopefully we'll get this out soon for you.

@puppetlabs-jenkins test this please

[puppetlabs-jenkins]

Refer to this link for build results (access rights to CI server needed):
http://jenkins-beaker.delivery.puppetlabs.net//job/qe_beaker-rspec_btc-intn-sys/44/

[kevpl]

@puppetlabs-jenkins retest this please

[puppetlabs-jenkins]

Refer to this link for build results (access rights to CI server needed):
http://jenkins-beaker.delivery.puppetlabs.net//job/qe_beaker-rspec_btc-intn-sys/46/

[barrymw]

Not a problem @kevpl, I have been running with my mods up until now. Thanks for getting around to merging though!

[kevpl]

@barrymw np! I'm pretty confident that your changes are fine & that we're having issues that we should resolve with our automation & shouldn't block merging on them. Since that's the case, I'll merge this now. Thanks for the addition!