Allowing partial specialization of the convert struct

[BenjaminNavarro]

This is done by adding a second, defaulted, template parameter that can
be used in conjunction of std::enable_if

An example of usage is added to the tutorial

The code and example has been taken from #310

[BenjaminNavarro]

How should I approach the testing of this? It's easy to check that it works for the correct subset of types but not that it doesn't for the others since it would create a compilation error

[BenjaminNavarro]

I added a small test to make sure that the feature works and to potentially avoid future regressions.

Let me know if you require more testing

[BenjaminNavarro]

@jbeder Could you review this PR when you have time?

[dota17]

Even though it's a test case, I think the constructors are needed, like #310. It is also convenient for later usage.

[BenjaminNavarro]

Added

[dota17]

How about updating the Tutorial.md at the same time. And use this in testcase.

  A a(1);
  B b(1, 2);

[BenjaminNavarro]

I updated the test cases to use the constructors and added them in the tutorial.
However the tutorial doesn't make use of them

[pettni]

I am also interested in getting this merged, was actually about to submit the same merge request myself. This change would facilitate using yaml-cpp in conjunction with reflection tools such as boost::hana and boost::fusion, and I don't see any drawbacks of adding the second default template argument.

Stackoverflow discussion that discusses the issue of not being able to use SFINAE without a second template paramter.

[yuriy1337]

Is there a plan for getting this merged. I am also interested in this functionality.

[BenjaminNavarro]

Is there any hope to see this merged eventually? It would be a life saver