Add options to mod info

[genebean]

While using this wonderful module at work today I ran into an issue where there was no way to modify the info.conf file created by include apache::mod::info The purpose of this PR is to add that functionality.

[genebean]

The test below are identical in syntax as far as I can see yet the FreeBSD one fails... not sure what's going on here and would love some help. The line general_info_specs() refers to the block of code containing all the tests.

  context 'On a Debian OS' do
    let :facts do
      {
        :osfamily               => 'Debian',
        :operatingsystemrelease => '6',
        :concat_basedir         => '/dne',
      }
    end

    # Load the more generic tests for this context
    general_info_specs()

    it { should contain_file('info.conf').with({
      :ensure => 'file',
      :path   => '/etc/apache2/mods-available/info.conf',
    } ) }
    it { should contain_file('info.conf symlink').with({
      :ensure => 'link',
      :path   => '/etc/apache2/mods-enabled/info.conf',
    } ) }
  end

  context 'on a RedHat OS' do
    let :facts do
      {
        :osfamily               => 'RedHat',
        :operatingsystemrelease => '6',
        :concat_basedir         => '/dne',
      }
    end

    # Load the more generic tests for this context
    general_info_specs()

    it { should contain_file('info.conf').with({
      :ensure => 'file',
      :path   => '/etc/httpd/conf.d/info.conf',
      } ) 
    }
  end

  context 'On a FreeBSD OS' do
    let :facts do
      {
        :osfamily               => 'FreeBSD',
        :operatingsystemrelease => '9',
        :concat_basedir         => '/dne',
      }
    end

    # Load the more generic tests for this context
    general_info_specs()

    it { should contain_file('info.conf').with({
      :ensure => 'file',
      :path   => '/usr/local/etc/apache22/Modules/info.conf',
    } ) }
  end

[igalic]

This line might be the reason why it fails on FreeBSD

[genebean]

@igalic I'll look at that more this afternoon but was wondering if you would elaborate on why the inclusion of mod_info would be causing this?

[igalic]

This is the only place this "extra" include actually happens.

Generally, multiple includes don't matter, however, in a later test, you're passing doing a resource definition to test a specific set of parameters.

There are two ways to fix this: Either, on FreeBSD you modify the default_mods in your test, or in general.

For the latter, I would create a Jira issue first.

[genebean]

Thanks! I'll investigate this today.

[igalic]

IIRC, the main reason why we have different set of default_mods for different platforms is so that we actually conform to their expected defaults.

[igalic]

@genebean any news on this? Can you rebase it as a starter?

[apenney]

^^ any news on the rebase?

[genebean]

Sorry, school and work got in the way of working on this. Going to attempt doing the rebase tonight and / or tomorrow. Actually pulling new code after posting this update.

[genebean]

Rebasing done, now on to trying to fix the duplicate declaration error.

[genebean]

Thanks for finding that line in the default mods list @igalic! I updated my testing to not include the default mods and all tests related to mod_info are now passing.

It seems some other modules are the cause of the Puppet 3.5 tests failing as all mod_info test pass there too.

[genebean]

@apenney - this was the first time I've had a need to do a rebase... if you happen to have any tips for how i could have improved on what I did here I welcome them.

[blkperl]

Github has some good documentation on how to rebase.

https://help.github.com/articles/about-git-rebase

Generally you would want something like

git rebase --interactive HEAD~18

but since you have some merge commits in there you will want to get rid of them first.

# or whatever origin has the github remote
git fetch upstream                 

# this will put your commit on top and remove the merge commits
git rebase upstream/master   

# now rebase using the number of commits
# choose the fixup option to squash all the commits, delete all the commit msgs and write a new 
# commit message that describes what this pr does
git rebase --interactive HEAD~<numofcommits> 

[genebean]

so, in this case since the code is passing builds, do I need to clean up the history or is it ready to merge?

[genebean]

Sorry @blkperl, reread your post and realized you were telling me to clean stuff up. I have squashed all commits into one. The new commit is genebean/puppetlabs-apache@28d2371

[genebean]

Just wanted to follow up and see what's next here.

[blkperl]

Any idea why the tests are failing?

[genebean]

Stuff unrelated to mod_info on Puppet 3.5. I see failures in the following

• apache::mod::negotiation
• apache::mod::php
• apache::mod::worker
• apache::fastcgi::server

[genebean]

FWIW, I see from the build that just ran (https://travis-ci.org/puppetlabs/puppetlabs-apache/builds/29471380) that the same 4 things are failing but only on Puppet 3.5.

[igalic]

This is just because of strict variables, @apenney is working super hard to fix that with puppet_facts

[genebean]

Thanks for the info @igalic, I appreciate it.