Puppet users - Apr 2012 - Facter not initialized in rspec in puppetlabs-stdlib?

I have been working on adding some predicate functions to puppetlabs-stdlib 
that test things based on ''interfaces'' and related facts.

The functions themselves seem to work fine in my manifest-based smoketests, 
but I am running into trouble that may or may not be due to my ignorance 
about rspec or Facter. Something needs to be done to initialize Facter 
because lookupvar(''interfaces'') just returns :undefined. I
have not found
any examples in the existing rspec tests; getvar_spec.rb seems like it 
should have something but the tests it does are pretty basic.

You can see some of what I''ve been trying at 
https://github.com/wcooley/puppetlabs-stdlib/blob/master/spec/unit/puppet/parser/functions/has_interface_with_spec.rb

Is there something obvious I''m missing?

Wil

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/puppet-users/-/XBJE-12hhCoJ.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Tue, Apr 10, 2012 at 11:00 PM, Wil Cooley <wilcooley@gmail.com>
wrote:
> I have been working on adding some predicate functions to puppetlabs-stdlib
> that test things based on ''interfaces'' and related facts.
Sweet!
> The functions themselves seem to work fine in my manifest-based smoketests,
> but I am running into trouble that may or may not be due to my ignorance
> about rspec or Facter. Something needs to be done to initialize Facter
> because lookupvar(''interfaces'') just returns :undefined.
I have not found
> any examples in the existing rspec tests; getvar_spec.rb seems like it
> should have something but the tests it does are pretty basic.
In general, you should assume the spec_helper for stdlib will
initialize facter for you.  In fact, you don''t need to do require
''facter'' in your example specification.
> You can see some of what I''ve been trying at
>
https://github.com/wcooley/puppetlabs-stdlib/blob/master/spec/unit/puppet/parser/functions/has_interface_with_spec.rb
Cool, I went ahead and pulled down your branch.  If I might make a
suggestion, I notice you''re developing on the "master branch"
and you
have a number of commits that have not actually been merged into the
"master" branch of the puppetlabs repository.

This situation of having two different branches, both named master, is
confusing and makes it a bit difficult to work with.

When using Git, creating new branches is almost effortless.  It''s just
two commands.

When you start hacking please create a topic branch:

# Create a branch indicating this is a new feature based on master
having something to do with an interface function
git branch feature/master/has_interface_with_function origin/master
# Checkout the branch
git checkout feature/master/has_interface_with_spec
# Hack away

It''s also fairly easy to rename your current topic branch which is
named "master"

# Rename your local master branch to a topic branch
git branch rename master  feature/master/has_interface_with_function

# Push the new name to your github repository
git push wcooley
feature/master/has_interface_with_function:feature/master/has_interface_with_function

# Set your local topic branch to track the published copy of it.
git branch --set-upstream feature/master/has_interface_with_function
wcooley/feature/master/has_interface_with_function

# Re-synchronize the Puppet Labs master branch (I assume origin is
git://github.com/puppetlabs/puppetlabs-stdlib.git)
git fetch origin

# Create your local master tracking Puppet Labs master  (These should
always be in sync.  Don''t commit on this branch...)
git branch master --track origin/master

# Replace your own master branch on Github with the Puppet Labs master
branch (because it''s not actually master right now)
git push wcooley +origin/master:master

# Finally, work in your topic branch
git co feature/master/has_interface_with_function
> Is there something obvious I''m missing?
The setup looks to be quite strange.  You shouldn''t ever need to
define a get_scope method.  I recommend looking at the
validate_re_spec.rb for a good example.

I''m also re-working your topic branch in my own if you''d like
to take
a look at the changes.

-Jeff

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

Wil,

I went ahead and re-worked the example specification.

Please take a look at the commit message and the way I''m setting an
expectation that mocks the behavior of lookupvar for both Mac OS X and
Linux.

https://github.com/jeffmccune/puppetlabs-stdlib/tree/feature/master/has_interface_with

$ rspec --format d spec/unit/puppet/parser/functions/has_interface_with_spec.rb

function_has_interface_with
  On Mac OS X Systems
    should have loopback (lo0)
    should not have loopback (lo)
  On Linux Systems
    should have loopback (lo)
    should not have loopback (lo0)

Finished in 0.04262 seconds
4 examples, 0 failures

Please feel free to reply here if you have any questions.  We might
also want to consider moving this thread to puppet-dev since it''s
developing a new function.

Hope this helps,
-Jeff

On Wed, Apr 11, 2012 at 9:03 AM, Jeff McCune <jeff@puppetlabs.com>
wrote:
> On Tue, Apr 10, 2012 at 11:00 PM, Wil Cooley <wilcooley@gmail.com>
wrote:
>> I have been working on adding some predicate functions to
puppetlabs-stdlib
>> that test things based on ''interfaces'' and related
facts.
>
> Sweet!
>
>> The functions themselves seem to work fine in my manifest-based
smoketests,
>> but I am running into trouble that may or may not be due to my
ignorance
>> about rspec or Facter. Something needs to be done to initialize Facter
>> because lookupvar(''interfaces'') just returns
:undefined. I have not found
>> any examples in the existing rspec tests; getvar_spec.rb seems like it
>> should have something but the tests it does are pretty basic.
>
> In general, you should assume the spec_helper for stdlib will
> initialize facter for you.  In fact, you don''t need to do require
> ''facter'' in your example specification.
>
>> You can see some of what I''ve been trying at
>>
https://github.com/wcooley/puppetlabs-stdlib/blob/master/spec/unit/puppet/parser/functions/has_interface_with_spec.rb
>
> Cool, I went ahead and pulled down your branch.  If I might make a
> suggestion, I notice you''re developing on the "master
branch" and you
> have a number of commits that have not actually been merged into the
> "master" branch of the puppetlabs repository.
>
> This situation of having two different branches, both named master, is
> confusing and makes it a bit difficult to work with.
>
> When using Git, creating new branches is almost effortless.  It''s
just
> two commands.
>
> When you start hacking please create a topic branch:
>
> # Create a branch indicating this is a new feature based on master
> having something to do with an interface function
> git branch feature/master/has_interface_with_function origin/master
> # Checkout the branch
> git checkout feature/master/has_interface_with_spec
> # Hack away
>
> It''s also fairly easy to rename your current topic branch which is
> named "master"
>
> # Rename your local master branch to a topic branch
> git branch rename master  feature/master/has_interface_with_function
>
> # Push the new name to your github repository
> git push wcooley
>
feature/master/has_interface_with_function:feature/master/has_interface_with_function
>
> # Set your local topic branch to track the published copy of it.
> git branch --set-upstream feature/master/has_interface_with_function
> wcooley/feature/master/has_interface_with_function
>
> # Re-synchronize the Puppet Labs master branch (I assume origin is
> git://github.com/puppetlabs/puppetlabs-stdlib.git)
> git fetch origin
>
> # Create your local master tracking Puppet Labs master  (These should
> always be in sync.  Don''t commit on this branch...)
> git branch master --track origin/master
>
> # Replace your own master branch on Github with the Puppet Labs master
> branch (because it''s not actually master right now)
> git push wcooley +origin/master:master
>
> # Finally, work in your topic branch
> git co feature/master/has_interface_with_function
>
>> Is there something obvious I''m missing?
>
> The setup looks to be quite strange.  You shouldn''t ever need to
> define a get_scope method.  I recommend looking at the
> validate_re_spec.rb for a good example.
>
> I''m also re-working your topic branch in my own if you''d
like to take
> a look at the changes.
>
> -Jeff
-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Wed, Apr 11, 2012 at 9:28 AM, Jeff McCune <jeff@puppetlabs.com>
wrote:
> Wil,
>
> I went ahead and re-worked the example specification.
>
> Please take a look at the commit message and the way I''m setting
an
> expectation that mocks the behavior of lookupvar for both Mac OS X and
> Linux.
>
>
https://github.com/jeffmccune/puppetlabs-stdlib/tree/feature/master/has_interface_with
Forgot to mention, the most recent commit in that branch is where I
re-worked the spec you asked about.

I''ve found it''s better to link to branches instead of commits
because
I often re-write history and commits are absolute references that
might go away inside of a branch whereas a branch is a relative
reference that can be updated frequently.

The actual commit I was referring to is:
https://github.com/jeffmccune/puppetlabs-stdlib/commit/dac3c508659bfc7d034509d3860afe5291aed722

Hope this helps,
-Jeff

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Wednesday, April 11, 2012 9:03:22 AM UTC-7, Jeff McCune
wrote:
>
> On Tue, Apr 10, 2012 at 11:00 PM, Wil Cooley <wilcooley@gmail.com>
wrote:
>
> In general, you should assume the spec_helper for stdlib will
> initialize facter for you.  In fact, you don''t need to do require
> ''facter'' in your example specification.
>
> > You can see some of what I''ve been trying at
> > 
>
https://github.com/wcooley/puppetlabs-stdlib/blob/master/spec/unit/puppet/parser/functions/has_interface_with_spec.rb
>
> Cool, I went ahead and pulled down your branch.  If I might make a
> suggestion, I notice you''re developing on the "master
branch" and you
> have a number of commits that have not actually been merged into the
> "master" branch of the puppetlabs repository.
>
Yeah, usually I do make branches; I think I mistakenly committed to master 
and then did not want to take the time to figure out how to move it on to a 
branch; your instructions below were very helpful. It would''ve taken me
some hours between re-reading Pro Git and searching stackoverflow to figure 
it out on my own.
 
> The setup looks to be quite strange.  You shouldn''t ever need to
define a
> get_scope method.  I recommend looking at the validate_re_spec.rb for a 
> good example.
>
It is strange; it''s the result of an hour or two of thrashing about, 
throwing bits at the wall and trying to see if anything stuck. Toss in some 
frustrations with gem and getting a working rspec environment and it''s
a
really bad dish.
 
> I''m also re-working your topic branch in my own if you''d
like to take
> a look at the changes.
>
Thanks;  I think I was expecting to be able to get facts from the live 
system, which is obviously inferior to mocking the facts that should be 
expected. I''ve merged in your branch and looked at your reworked
version; I
think it makes sense and I see how to proceed with adding the rest of my 
tests.

Unfortunately, I am unable to run any of the rspec tests--not just in 
has_interface_with_spec.rb, but on any of the specs. They fail with:

     Failure/Error: Unable to find matching line from backtrace
     NoMethodError:
       undefined method `initialize_everything_for_tests'' for 
#<Puppet::Util::Settings:0x2b7d43d59940>

I''ve got 2.7.13, RPMs built yesterday, both puppet and puppet-master 
installed:

$ rpm -qi puppet
Name        : puppet                       Relocations: (not relocatable)
Version     : 2.7.13                            Vendor: (none)
Release     : 1.el5                         Build Date: Tue 10 Apr 2012 
01:52:57 PM PDT
Install Date: Wed 11 Apr 2012 09:46:53 PM PDT      Build Host: 
rpm-builder.puppetlabs.lan
...

ISTR seeing something about this method in the commit logs recently...

Also, go ahead and reply to puppet-dev if you wish; I''m trying to use
the
Google Groups interface and it does not appear to let me direct my response 
to a different group.

Wil

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/puppet-users/-/Y1_YL3HbfVgJ.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Wed, Apr 11, 2012 at 9:59 PM, Wil Cooley <wilcooley@gmail.com>
wrote:
> Unfortunately, I am unable to run any of the rspec tests--not just in
> has_interface_with_spec.rb, but on any of the specs. They fail with:
>
>      Failure/Error: Unable to find matching line from backtrace
>      NoMethodError:
>        undefined method `initialize_everything_for_tests'' for
> #<Puppet::Util::Settings:0x2b7d43d59940>
>
> I''ve got 2.7.13, RPMs built yesterday, both puppet and
puppet-master
> installed:
Ah, this is a problem our platform team is actively working on right now.

The problem is that modules like stdlib need Puppet to be initialized
in a manner suitable for testing.  Setting things like confdir and
what not.  This means different actions for different versions of
Puppet, and stdlib works all the way back through 2.6.

In any event, if you run against Puppet 2.7.x from source, this method
should already be added to Puppet.

We should also have stdlib fixed up shortly.

If you''d like to actively work on this, please feel free to hop into
#puppet and or #puppet-dev on freenode and ping cprice or myself.

More background information on this problem is located at:
http://projects.puppetlabs.com/issues/13693

Hope this helps,
-Jeff

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Thursday, April 12, 2012 10:06:25 AM UTC-7, Jeff McCune
wrote:
>
> The problem is that modules like stdlib need Puppet to be initialized
> in a manner suitable for testing.  Setting things like confdir and
> what not.  This means different actions for different versions of
> Puppet, and stdlib works all the way back through 2.6.
>
I took the expedient way out and commented out the requires from 
spec_helper. :-\ My tests pass now, at least.

I have now added tests for the has_ip_address and has_ip_network, which are 
more porcelain than has_interface_with.

https://github.com/wcooley/puppetlabs-stdlib/compare/puppetlabs:master...wcooley:has_interface_with_function

If you like it, I will open a ticket and submit a pull request.

Wil
 

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/puppet-users/-/q8neT0W4zUcJ.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Sat, Apr 14, 2012 at 8:34 PM, Wil Cooley <wilcooley@gmail.com>
wrote:
> On Thursday, April 12, 2012 10:06:25 AM UTC-7, Jeff McCune wrote:
>>
>> The problem is that modules like stdlib need Puppet to be initialized
>> in a manner suitable for testing.  Setting things like confdir and
>> what not.  This means different actions for different versions of
>> Puppet, and stdlib works all the way back through 2.6.
>
> I took the expedient way out and commented out the requires from
> spec_helper. :-\ My tests pass now, at least.
>
> I have now added tests for the has_ip_address and has_ip_network, which are
> more porcelain than has_interface_with.
>
>
https://github.com/wcooley/puppetlabs-stdlib/compare/puppetlabs:master...wcooley:has_interface_with_function
>
> If you like it, I will open a ticket and submit a pull request.
Yes, please file a ticket at:

http://projects.puppetlabs.com/projects/stdlib

You''ll need to register an account if you don''t already have
one.

If it''s a somewhat substantial patch could you also please sign the CLA
at:
https://projects.puppetlabs.com/contributor_licenses/sign

Thanks,
-Jeff McCune

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.

On Apr 16, 2:01 pm, Jeff McCune <j...@puppetlabs.com>
wrote:
> On Sat, Apr 14, 2012 at 8:34 PM, Wil Cooley <wilcoo...@gmail.com>
wrote:
> > On Thursday, April 12, 2012 10:06:25 AM UTC-7, Jeff McCune wrote:
>
> >> The problem is that modules like stdlib need Puppet to be
initialized
> >> in a manner suitable for testing.  Setting things like confdir and
> >> what not.  This means different actions for different versions of
> >> Puppet, and stdlib works all the way back through 2.6.
>
> > I took the expedient way out and commented out the requires from
> > spec_helper. :-\ My tests pass now, at least.
>
> > I have now added tests for the has_ip_address and has_ip_network,
which are
> > more porcelain than has_interface_with.
>
>
>https://github.com/wcooley/puppetlabs-stdlib/compare/puppetlabs:maste...
>
> > If you like it, I will open a ticket and submit a pull request.
>
> Yes, please file a ticket at:
>
> http://projects.puppetlabs.com/projects/stdlib
>
> You''ll need to register an account if you don''t already
have one.
>
> If it''s a somewhat substantial patch could you also please sign
the CLA at:https://projects.puppetlabs.com/contributor_licenses/sign
I thought I responded that I had opened a ticket and submitted a pull
request, but I guess I didn''t.

Anyway, here''s the ticket:

https://projects.puppetlabs.com/issues/13974

And here''s the pull request:

https://github.com/puppetlabs/puppetlabs-stdlib/pull/64

Would someone with access fix the typo in the subject of the ticket?
Change "not to use" to "not easy to use".

Thanks!
Wil

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.