Updates to work with Ruby 2.4.0 #13104

Fryguy · 2016-12-10T04:18:04Z

This is one step to ruby 2.4, here is an outline of the steps: #14446

Known issues:

RubyParser didn't support Ruby 2.4 immediately, and now does in 3.9.0 and here, with changelog here. RubyParser tends to lag, so instead I found RipperRubyParser which is a "drop-in replacement for RubyParser using Ripper". Since Ripper comes with Ruby it will always be up-to-date.
Expand symbol to proc in a factory due to a ruby 2.4.0 bug …
https://bugs.ruby-lang.org/issues/13074
via
upgrading from 4.7.0 to 4.8.0 breaks symbol to proc callbacks thoughtbot/factory_bot#980
- "When executing instance_exec with symbol.to_proc, it ignores first
  argument"
- Fixed in ruby 2.4.1
Ruby 2.4 unifies Fixnum and Bignum into Integer, the former classes are deprecated
Monkey patch webmock for now, may need to re-record cassettes with new webmock supporting ruby 2.4: Added Update to webmock 2.3.1+ and VCR 3.0.2+ #14270 to upgrade webmock and remove this temporary monkey patch.

NOTE: We are not changing the required ruby version in the Gemfile nor the one tested by travis in the .travis.yml in this PR. These changes will work with ruby 2.4 and 2.3 and probably even 2.2.

cc @jrafanie

jrafanie · 2016-12-14T21:18:45Z

wow, so many warnings... for future me, these were the test failures:

  1) Metric as vmware Testing CPU % virtual cols with existing performance data should return the correct values for Vm realtime
     Failure/Error: expect(perf.v_pct_cpu_wait_delta_summation).to eq(135.5)
     
       expected: 135.5
            got: 135.4
     
       (compared using ==)
     # ./spec/models/metric_spec.rb:923:in `block (4 levels) in <top (required)>'

  2) ManageIQ::Providers::AnsibleTower::ConfigurationManager::ConfigurationScript belongs_to the Ansible Tower manager
     Failure/Error: let(:manager)      { FactoryGirl.create(:configuration_manager_ansible_tower, :provider, :configuration_script) }
     
     NoMethodError:
       undefined method `create_provider' for #<FactoryGirl::SyntaxRunner:0x0000003c92ea78>
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:8:in `block (2 levels) in <top (required)>'
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:11:in `block (2 levels) in <top (required)>'

  3) ManageIQ::Providers::AnsibleTower::ConfigurationManager::ConfigurationScript#run launches the referenced ansible job template
     Failure/Error: let(:manager)      { FactoryGirl.create(:configuration_manager_ansible_tower, :provider, :configuration_script) }
     
     NoMethodError:
       undefined method `create_provider' for #<FactoryGirl::SyntaxRunner:0x0000003c92ea78>
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:8:in `block (2 levels) in <top (required)>'
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:24:in `block (3 levels) in <top (required)>'

  4) ManageIQ::Providers::AnsibleTower::ConfigurationManager::ConfigurationScript#run accepts different variables to launch a job template against
     Failure/Error: let(:manager)      { FactoryGirl.create(:configuration_manager_ansible_tower, :provider, :configuration_script) }
     
     NoMethodError:
       undefined method `create_provider' for #<FactoryGirl::SyntaxRunner:0x0000003c92ea78>
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:8:in `block (2 levels) in <top (required)>'
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:30:in `block (3 levels) in <top (required)>'

  5) ManageIQ::Providers::AnsibleTower::ConfigurationManager::ConfigurationScript#merge_extra_vars merges an empty internal hash and a hash to send out to the tower gem
     Failure/Error: let(:manager)      { FactoryGirl.create(:configuration_manager_ansible_tower, :provider, :configuration_script) }
     
     NoMethodError:
       undefined method `create_provider' for #<FactoryGirl::SyntaxRunner:0x0000003c92ea78>
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:8:in `block (2 levels) in <top (required)>'
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:52:in `block (3 levels) in <top (required)>'

  6) ManageIQ::Providers::AnsibleTower::ConfigurationManager::ConfigurationScript#merge_extra_vars merges all empty arguments to send out to the tower gem
     Failure/Error: let(:manager)      { FactoryGirl.create(:configuration_manager_ansible_tower, :provider, :configuration_script) }
     
     NoMethodError:
       undefined method `create_provider' for #<FactoryGirl::SyntaxRunner:0x0000003c92ea78>
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:8:in `block (2 levels) in <top (required)>'
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:60:in `block (3 levels) in <top (required)>'

  7) ManageIQ::Providers::AnsibleTower::ConfigurationManager::ConfigurationScript#merge_extra_vars merges an internal hash and an empty hash to send out to the tower gem
     Failure/Error: let(:manager)      { FactoryGirl.create(:configuration_manager_ansible_tower, :provider, :configuration_script) }
     
     NoMethodError:
       undefined method `create_provider' for #<FactoryGirl::SyntaxRunner:0x0000003c92ea78>
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:8:in `block (2 levels) in <top (required)>'
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:44:in `block (3 levels) in <top (required)>'

  8) ManageIQ::Providers::AnsibleTower::ConfigurationManager::ConfigurationScript#merge_extra_vars merges internal and external hashes to send out to the tower gem
     Failure/Error: let(:manager)      { FactoryGirl.create(:configuration_manager_ansible_tower, :provider, :configuration_script) }
     
     NoMethodError:
       undefined method `create_provider' for #<FactoryGirl::SyntaxRunner:0x0000003c92ea78>
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:8:in `block (2 levels) in <top (required)>'
     # ./spec/models/manageiq/providers/ansible_tower/configuration_manager/configuration_script_spec.rb:36:in `block (3 levels) in <top (required)>'

  9) MiqServer RhnMirror #resync_rhn_mirror
     Failure/Error: FileUtils.rm_f(APACHE_MIRROR_CONF_FILE)
     
       FileUtils received :rm with unexpected arguments
         expected: ("/etc/httpd/conf.d/manageiq-https-mirror.conf", {:force=>true})
              got: ("/etc/httpd/conf.d/manageiq-https-mirror.conf", {:force=>true, :noop=>nil, :verbose=>nil})
       Diff:
       @@ -1,2 +1,3 @@
       -["/etc/httpd/conf.d/manageiq-https-mirror.conf", {:force=>true}]
       +["/etc/httpd/conf.d/manageiq-https-mirror.conf",
       + {:force=>true, :noop=>nil, :verbose=>nil}]
       
     # ./app/models/miq_server/rhn_mirror.rb:82:in `create_apache_vhost'
     # ./app/models/miq_server/rhn_mirror.rb:40:in `configure_rhn_mirror_server'
     # ./app/models/miq_server/rhn_mirror.rb:27:in `resync_rhn_mirror'
     # ./spec/models/miq_server/rhn_mirror_spec.rb:45:in `block (3 levels) in <top (required)>'

  10) CustomAttribute returns the value type of Fixnum custom attributes
      Failure/Error: expect(int_custom_attribute.value_type).to eq(:fixnum)
      
        expected: :fixnum
             got: :integer
      
        (compared using ==)
      
        Diff:
        @@ -1,2 +1,2 @@
        -:fixnum
        +:integer
        
      # ./spec/models/custom_attribute_spec.rb:15:in `block (2 levels) in <top (required)>'

Finished in 20 minutes 39 seconds (files took 46.46 seconds to load)
7162 examples, 10 failures, 8 pending

Fryguy · 2016-12-20T20:21:07Z

@miq-bot add_label performance

bdunne · 2017-01-26T16:23:00Z

@Fryguy I updated manageiq-gems-pending with ManageIQ/manageiq-gems-pending#47 and tried adding 2.4, but there were many failures and deprecations, so I removed that from the PR temporarily

jrafanie · 2017-02-02T22:45:28Z

spec/support/webmock_ruby24_bridge.rb

+class StubSocket
+  def close
+  end
+end


We can either upgrade webmock or do this for now. I tried upgrading it locally and it looks at the very least that we'd need to re-record our vcr cassettes. I'm not sure if that's all, so I went the easy route here in hopes we can upgrade in a followup PR.

note, @cben I'd love your feedback on the failures in #14270 since it looks like much of the failures are VCR cassettes you may know how to re-record.

jrafanie · 2017-02-02T22:46:11Z

spec/models/miq_server/rhn_mirror_spec.rb

@@ -34,7 +34,7 @@

      expect(FileUtils).to receive(:mkdir_p).with("/repo/mirror")
      expect(MiqApache::Conf).to receive(:create_conf_file).once.and_return(true)
-      expect(FileUtils).to receive(:rm).with("/etc/httpd/conf.d/manageiq-https-mirror.conf", :force => true)
+      expect(FileUtils).to receive(:rm).with("/etc/httpd/conf.d/manageiq-https-mirror.conf", hash_including(:force => true))


See commit, it's basically changed from an options hash to kwargs so the expectations on arguments is different now.

jrafanie · 2017-02-02T22:48:34Z

spec/models/custom_attribute_spec.rb

-
-  it "returns the value type of Fixnum custom attributes" do
-    expect(int_custom_attribute.value_type).to eq(:fixnum)
-  end


See the commit regarding custom attributes, we should really find out if we're actually using value_type because it's really brittle to use class names instead of using interfaces such as expecting fixnum/bignum/integer types to respond to comparison operators. CC @alongoldboim (see the commit for more details)

I was pretty sure that we use integer for these

cc @alongoldboim can you review this commit? It's not clear why we're using hardcoded ruby class names here. Do you know if it's needed?

cc @zgalor do you know if/how CustomAttribute#value_type is being used?
It's defined as serialized_value.class.to_s.downcase.to_sym :-(
I doubt removing the spec is enough. We likely want either backward compat (return :fixnum on 2.4), or forward compat (switch to :integer on <=2.3), or remove the functionality.

API is only accepting (on POST?) a whitelist for field_type: https://github.com/ManageIQ/manageiq/search?utf8=%E2%9C%93&q=ALLOWED_API_VALUE_TYPES
https://github.com/ManageIQ/manageiq/blob/caa745c49/app/controllers/api/providers_controller.rb#L90
Not sure if/how field_type interacts with value_type and serialized_value. Do we expose value_type on GET?

Sorry, I was lazy, I didn't put the commit details (which keeps changing the commit sha1 when I rebase/force push), so I'll put it here. Note, I totally agree with @cben here, the test removal is probably not enough. We need to remove usage that cared if it was Fixnum vs Bignum:

CustomAttribute#value_type returns a symbol for a ruby core class, a
hardcoded value instead of the class's behavior.
Do we use this? Why are we returning a value instead of caring about
behavior of the old Fixnum and Integer? If we don't use this, we
should remove the whole method and it's tests added in:
4687394

The PR that added this was:
#11000

Adding CustomAttribute to provider UI/API was added in:
#10897

kbrock

kbrock · 2017-02-02T23:11:27Z

spec/models/custom_attribute_spec.rb

-
-  it "returns the value type of Fixnum custom attributes" do
-    expect(int_custom_attribute.value_type).to eq(:fixnum)
-  end


I was pretty sure that we use integer for these

kbrock · 2017-02-02T23:12:04Z

spec/factories/ext_management_system.rb

@@ -327,7 +327,9 @@
          :parent  => :configuration_manager

  trait(:provider) do
-    after(:build, &:create_provider)
+    after(:build) do |x|


huh. any reasoning behind this one?

Sounds like the commit message specifically explains the bug: ~~d94a400~~ ~~296b525~~

Edit: Update link do to rebase.

Updated again due to rebase e0825a0ff7078641680b1a42a82db8e0875f6dd0

kbrock · 2017-02-02T23:17:03Z

Could you double check [travis]? it seems quite vocal

kbrock · 2017-02-24T21:59:56Z

ugh, is https://rubygems.org/gems/ezcrypto abandoned? Don't know if https://github.com/pelle/ezcrypto is the new owner - 8 hours old

jrafanie · 2017-02-24T22:09:04Z

ugh, is https://rubygems.org/gems/ezcrypto abandoned? Don't know if https://github.com/pelle/ezcrypto is the new owner - 8 hours old

Yeah, I don't know @kbrock, I don't know if we can just swap our usage of ezcrypto out for something else at this point. I wish we could.

kbrock · 2017-02-25T04:28:10Z

oops s/hours/years - ezcrypto hasn't been modified in 8 years

But yes pelle is the owner of the gem at rubygems. (and luckily he is still active and has modified some ruby code recently - just not this gem)

kbrock · 2017-02-25T04:37:57Z

We may want to give them a PR with a .travis.yml. Ping me if I drop the ball on this one

miq-bot · 2017-03-10T14:27:27Z

This pull request is not mergeable. Please rebase and repush.

jrafanie · 2017-03-21T23:15:23Z

@Fryguy @chessbyte I moved the checklist to #14446

chrisarcand · 2017-03-22T15:57:55Z

Are all of the corresponding gems (more_core, linux_admin, etc) testing against 2.4? If not, do you think we should update them all ahead of time?

Seems wise and inexpensive to require beforehand.

Set Travis to Ruby 2.4 and miq-bot to Ruby 2.3 (mirror what we do now)

Why miq-bot to 2.3 instead of 2.4?

announce the end of support for Ruby 2.2 when we merge, and only support Ruby 2.3+

🎉

Fryguy · 2017-03-22T18:34:43Z

Why miq-bot to 2.3 instead of 2.4?

This essentially allows a syntax check against Ruby 2.3, so we can keep Ruby 2.3 as a minimum, without forcing us to run tests for both 2.3 and 2.4 in travis (which is expensive).

Copied from: ManageIQ/manageiq#13104 Ruby 2.4.0 removed the closed? check in the conditional in: s.close if !s.closed? Webmock was changed to add close to StubSocket along with another change.. ruby/ruby@f845a9e bblimke/webmock@8f2176a WebMock 2.3.1+ fixed the issue with ruby 2.4.0 by adding StubSocket#close.

kbrock · 2017-03-28T14:10:48Z

You know, the parser is the only real change.
If we are comfortable with that change, then let's just merge this now.
Or if we are not comfortable with that change, let's take out the minor bug fixes and have a long standing "please merge" PR with just that change.

jrafanie · 2017-03-29T21:30:36Z

You know, the parser is the only real change.
If we are comfortable with that change, then let's just merge this now.
Or if we are not comfortable with that change, let's take out the minor bug fixes and have a long standing "please merge" PR with just that change.

we can even monkey patch ruby parser or use master branch of ruby-parser (see the links in the description)

jrafanie · 2017-04-17T20:46:57Z

@Fryguy @bdunne I'd like to get this in since the providers checkout miq to run their tests. Now that fine is branched, is there any risk to merging this on master? If the ripper ruby parser change is scary, we can use the 3.9.0 version of ruby parser which seems to now support ruby 2.4 and future versions.

https://github.com/seattlerb/ruby_parser/blob/master/History.rdoc

bdunne

👍 LGTM

NickLaMuro

This is the only thing I noticed that hasn't been addressed, but even then, it isn't a blocker and I don't have a suggestion of my own. Looks good to me.

NickLaMuro · 2017-04-17T21:18:23Z

lib/miq_automation_engine/engine/miq_ae_engine/drb_remote_invoker.rb

-               .try(:instance_variable_get, '@holder')
-               .try(:instance_variable_get, '@keeper')
-      if RUBY_VERSION > "2.3.1"
-        warn "Remove me: #{__FILE__}:#{__LINE__} and my test. Ruby 2.3.2+ should not be creating timer threads."


Because you are removing this warn, does that mean you have given up on ever removing this hack? Or do you have some other method in mind for remembering to remove this when ruby 2.3 is no longer supported?

A spec that fails when the ruby version is higher than some version? (e.g.: 2.5)
(that way assuming that we stopped supporting 2.3)?

Interesting @NickLaMuro @kbrock, especially the 2.5 test. The problem is we usually support two ruby versions on master, so yeah, we can't remove this hack completely until ruby 2.3.0 and below are unsupported.

We can as of bundler 1.12+, add
ruby "~> 2.3", "< 2.4.1" to the Gemfile, but we can't programmatically access this to check if the current ruby is satisfied.

It would be nice to add a warn if Bundler.load.required_ruby.satisfied_by?(RUBY_VERSION) was a thing but the ruby requirement isn't exposed in the bundler runtime... See rubygems/bundler@b4a52c3. Maybe I'm missing something but I couldn't access it.

Would maybe some kind of .travis.yml parser that checks the supported ruby versions be a good way to handle this? Might be over-engineering things a bit, but something to consider so we aren't forgetting to remove this in the future. This of course assumes that the .travis.yml is consistent here and in downstream as a source of truth for supported ruby versions.

Otherwise I am at a loss on how we could do this.

In an ideal world, we would open a PR now to remove it, which would fail tests now but start passing as soon as .travis.yml tests only >2.3.1.
But something would have to periodically re-run the tests and notify someone once they start passing. Or directly "merge when green", measured in years

"An ideal world is left as an exercise to the reader" — Paul Graham, On Lisp

Silly question:
We're bumping to requiring ≥2.3 now that we'll support 2.4, no?
Can't we simply proclaim we support ≥2.3.2 and drop the hack?

miq-bot · 2017-05-01T17:57:12Z

This pull request is not mergeable. Please rebase and repush.

RubyParser lags Ruby and doesn't yet have Ruby 2.4 support. RipperRubyParser uses Ripper and so will always have the latest version.

ruby/ruby@baa43cd Fixes: expected: ("/etc/httpd/conf.d/manageiq-https-mirror.conf", {:force=>true}) got: ("/etc/httpd/conf.d/manageiq-https-mirror.conf", {:force=>true, :noop=>nil, :verbose=>nil})

Ruby 2.4.0 removed the closed? check in the conditional in: s.close if !s.closed? Webmock was changed to add close to StubSocket along with another change.. ruby/ruby@f845a9e bblimke/webmock@8f2176a WebMock 2.3.1+ fixed the issue with ruby 2.4.0 by adding StubSocket#close.

For now, make the test expectations different for ruby 2.4.0+ vs. prior rubies. CustomAttribute#value_type returns a symbol for a ruby core class, a hardcoded value instead of the class's behavior. Do we use this? Why are we returning a value instead of caring about behavior of the old Fixnum and Integer? If we don't use this, we should remove the whole method and it's tests added in: ManageIQ@4687394 The PR that added this was: ManageIQ#11000 Adding CustomAttribute to provider UI/API was added in: ManageIQ#10897

miq-bot · 2017-05-01T19:47:42Z

Checked commits Fryguy/manageiq@90c1d06~...56f5fc9 with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0
5 files checked, 0 offenses detected
Everything looks fine. 🏆

https://bugs.ruby-lang.org/issues/12342 has been fixed in 2.3.2+ and 2.4.0+ so we should only do this HACK for 2.3.1 and lower. Extracted from ManageIQ/manageiq#13104

Fryguy · 2017-05-01T20:37:30Z

Even though this is my PR, none of the code left is mine anymore 😆 , so I'm going to merge.

Fryguy assigned jrafanie Dec 10, 2016

Fryguy added wip ruby2.4 labels Dec 10, 2016

miq-bot added the dependencies label Dec 10, 2016

chrisarcand self-assigned this Dec 13, 2016

miq-bot added the performance label Dec 20, 2016

juliancheal mentioned this pull request Jan 3, 2017

[WIP] Testing Ruby 2.4.0 #13316

Closed

jrafanie force-pushed the ruby24 branch 3 times, most recently from b144ead to 0b6f7e0 Compare February 2, 2017 21:59

jrafanie reviewed Feb 2, 2017

View reviewed changes

kbrock approved these changes Feb 2, 2017

View reviewed changes

chrisarcand removed their assignment Feb 3, 2017

jrafanie force-pushed the ruby24 branch from 0b6f7e0 to fd77dd7 Compare February 23, 2017 21:59

jrafanie force-pushed the ruby24 branch from fd77dd7 to 82c7c8f Compare March 3, 2017 16:50

Fryguy mentioned this pull request Mar 8, 2017

Avoid calling $evm.backtrace over DRb to prevent DRb-level mutex locks #14239

Merged

miq-bot added the unmergeable label Mar 10, 2017

jrafanie force-pushed the ruby24 branch from 99190f6 to 4c15cea Compare March 10, 2017 14:28

miq-bot removed the unmergeable label Mar 10, 2017

jrafanie force-pushed the ruby24 branch from de93339 to 28dcd92 Compare March 24, 2017 15:39

jrafanie mentioned this pull request Mar 24, 2017

Test updated ruby 2.3.3 and new 2.4.1. ManageIQ/manageiq-providers-lenovo#36

Merged

This was referenced Mar 24, 2017

Add ruby 2.4 support ManageIQ/manageiq-gems-pending#101

Merged

Add ruby 2.4 support and bump 2.3 ManageIQ/manageiq-ui-classic#803

Merged

jrafanie force-pushed the ruby24 branch from 28dcd92 to ef9659d Compare March 27, 2017 14:14

bdunne approved these changes Apr 17, 2017

View reviewed changes

NickLaMuro approved these changes Apr 17, 2017

View reviewed changes

miq-bot added the unmergeable label May 1, 2017

Fryguy and others added 4 commits May 1, 2017 15:42

Replace RubyParser with RipperRubyParser

90c1d06

RubyParser lags Ruby and doesn't yet have Ruby 2.4 support. RipperRubyParser uses Ripper and so will always have the latest version.

The method signature looks different since it now uses kwargs.

ef992a1

ruby/ruby@baa43cd Fixes: expected: ("/etc/httpd/conf.d/manageiq-https-mirror.conf", {:force=>true}) got: ("/etc/httpd/conf.d/manageiq-https-mirror.conf", {:force=>true, :noop=>nil, :verbose=>nil})

jrafanie force-pushed the ruby24 branch from ef9659d to 56f5fc9 Compare May 1, 2017 19:44

miq-bot removed the unmergeable label May 1, 2017

jrafanie mentioned this pull request May 1, 2017

Kill the TimerIdConv thread only for ruby 2.3.1 and lower. ManageIQ/manageiq-automation_engine#1

Merged

Fryguy merged commit 7205128 into ManageIQ:master May 1, 2017

Fryguy added this to the Sprint 60 Ending May 8, 2017 milestone May 1, 2017

Fryguy deleted the ruby24 branch May 5, 2017 19:00

chargio mentioned this pull request Jun 14, 2017

improving requirements for updates in the development guide ManageIQ/manageiq-vagrant-dev#3

Merged

Updates to work with Ruby 2.4.0 #13104

Updates to work with Ruby 2.4.0 #13104

Conversation

Fryguy commented Dec 10, 2016 • edited by jrafanie Loading

jrafanie commented Dec 14, 2016 • edited Loading

Fryguy commented Dec 20, 2016

bdunne commented Jan 26, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kbrock left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

NickLaMuro Feb 2, 2017 • edited by jrafanie Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kbrock commented Feb 2, 2017

kbrock commented Feb 24, 2017

jrafanie commented Feb 24, 2017

kbrock commented Feb 25, 2017 • edited Loading

kbrock commented Feb 25, 2017

miq-bot commented Mar 10, 2017

jrafanie commented Mar 21, 2017

chrisarcand commented Mar 22, 2017

Fryguy commented Mar 22, 2017

kbrock commented Mar 28, 2017

jrafanie commented Mar 29, 2017

jrafanie commented Apr 17, 2017

bdunne left a comment

Choose a reason for hiding this comment

NickLaMuro left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

miq-bot commented May 1, 2017

miq-bot commented May 1, 2017

Fryguy commented May 1, 2017

Fryguy commented Dec 10, 2016 •

edited by jrafanie

Loading

jrafanie commented Dec 14, 2016 •

edited

Loading

NickLaMuro Feb 2, 2017 •

edited by jrafanie

Loading

kbrock commented Feb 25, 2017 •

edited

Loading