Add support for activesupport gem?

[mtho11]

I'm just reiterating @cfcosta question in irc about adding support for activesupportgem. It sounds like a good idea to me as it will make our code more rails idiomatic.

Thoughts?

[cfcosta]

I'm 👍 to that. It simplifies a lot some parts of code, and the tradeoff (having another maybe heavy dependency) is not that big imho.

[josejulio]

I also think we can benefit from having that dependency. 👍

[josejulio]

Although we might want to wait for #196 before refactoring anything.

[abonas]

I missed the beginning of the discussion - please provide some examples as to why this is beneficial to this particular gem.
Please also see an issue that we encountered in another gem which actually caused us to drop the dependency in active support.
ManageIQ/kubeclient#182

[cfcosta]

@abonas what's our target version of Ruby (and Rails)? Enforcing 2.2.2+ seems good to me, given that it's almost 2 years old.

[pilhuhn]

On 20 Mar 2017, at 0:30, Cainã Costa wrote: @abonas what's our target version of Ruby (and Rails)? Enforcing 2.2.2+ seems good to me, given that it's almost 2 years old.

I think that is fine.

[pilhuhn]

While it looks like this would work also in non-Rails environments, I want to make sure it indeed does.

I would hate to lose the ability to use our client gem outside of Rails.

[abonas]

On 20 Mar 2017, at 0:30, Cainã Costa wrote:
@abonas what's our target version of Ruby (and Rails)? Enforcing
2.2.2+ seems good to me, given that it's almost 2 years old.
I think that is fine.

Right now the gemspec says >2.0 . I don't have any objection to bump it, but in the current situation 'as is' adding activesupport will get us into an issue.
And I also would like to make sure as @pilhuhn that this will continue working in a non Rails env, although I am pretty sure it should.

[cfcosta]

@pilhuhn @abonas it works just fine, activesupport is not rails-bound at all (there are some stuff used by rails, but they are only bound in the railties gem). Also, we do not need to require everything, just the extensions we need, like requiring active_support/core_ext/blank if we need blank?. By doing both we avoid polluting the user namespace too much and still have the niceties for ourselves.

[josejulio]

This is what we currently test in travis.
We stopped testing against 2.0 and 2.1 in #165.

[abonas]

so lets bump the gemspec, especially if we don't test 2.0 and 2.1 anyway.
(if we don't test it, how can we support it?)

[josejulio]

Did we ever include this ?