Fix mounting APIs in route_param namespaces

[wojciechka]

This is based on #543 by @milgner and addresses issue #267

I have applied the changes on top of latest master and included test cases.

[coveralls]

Coverage increased (+2.0%) to 99.642% when pulling b29a8ed on wojciechka:fix267 into e9d9447 on ruby-grape:master.

[LeFnord]

many thanks for your work @wojciechka … please fix rubocop offences and I'm happyo to mörge it :)

[LeFnord]

@wojciechka … please wait, fixing of rubocop stuff is out of scope of this PR

[wojciechka]

@LeFnord I have just pushed fixes that clean up offences related to my changes.

As you have probably noticed, most of the issues are not related to the changes in code I have made.

But I also agree that I should not be introducing new offences.

[LeFnord]

thanks again @wojciechka … that's what I meant with out of scope ;)
please can you fix the rubocop version to 0.49.1, so all should be fine

[wojciechka]

@LeFnord sure, just pushed an update ; also, I am curious, wouldn't Gemfile.lock prevent those things from happening (as in an update in rubocop version triggering a lot of false positives)?

I usually commit Gemfile.lock to avoid package versions changing between deployments etc and I saw it explicitly removed.

It's not really related to this PR, but I am curious about it.

[coveralls]

Coverage increased (+2.0%) to 99.642% when pulling c5747cd on wojciechka:fix267 into e9d9447 on ruby-grape:master.

[wojciechka]

@LeFnord With the rubocop version fix it has passed all the tests now.

Anything pending on my side?

[LeFnord]

@wojciechka yhanks again :) … happy to mörge it

[rypit]

Also addresses issues similar to #610 & #508 which were happening for me in 0.27.3. Thanks for this 👍