Fix debugging output for nameless py::arg_v annotations

[jagerman]

This fixes a couple bugs with nameless py::arg_v (introduced in #634) annotations:

• the argument name was being used in debug mode without checking that it exists (which would result in the std::string construction throwing an exception for being invoked with a nullptr)
• the error output says "keyword arguments", but py::arg_v() can now also be used for positional argument defaults.
• the debugging output "in function named 'blah'" was overly verbose: changed it to just "in function 'blah'".

[dean0x7d]

What's the rationale for adding new tests to test_issues instead of to the relevant test_* files? It seems like the tests/issues split makes it harder to located all related tests.

[wjakob]

@dean0x7d: it's historical: the examples used to be actual examples, and it felt weird to move issue code into those files which were meant as additional demonstrations of certain binding features.

[wjakob]

(this was from a time when this project really wasn't all that well-documented)

[wjakob]

@dean0x7d: Ah, sorry -- I misunderstood (you were referring to @jagerman's commit). I thought that you were generally questioning the idea of putting tests into test_issues.cpp

[dean0x7d]

Ah, sorry -- I misunderstood (you were referring to @jagerman's commit). I thought that you were generally questioning the idea of putting tests into test_issues.cpp

A little bit of both I guess. I understand the historical reason for the split, but I'm just suggesting that it would be nicer to group by functionality going forward (for ease of maintenance) since these are definitely not examples anymore.

[jagerman]

No idea what happened to the last test--looks like a appveyor problem (the first three passed).

[wjakob]

I'll merge it -- let's see what happens with AppVeyor :)