Unicode escaping

[Teemperor]

Implements the \uXXXX and \uXXXX\uXXXX unicode escape mechanisms of JSON.

This branch has a problem with code coverage as there is one method which has a untestable case (line 2215).
This code is only reachable if

1. there is a bug somewhere else that calls the function with the wrong argument
2. we have access to the private method.

So we need a way to directly access the codePointToUTF8 method from the test header to get the line coverage back to 100%.

Any recommendation how we do that in a nice way?

[coveralls]

Coverage decreased (-0.39%) when pulling 6105ce5 on Teemperor:unicode_escaping into 0cd2ecf on nlohmann:master.

[nlohmann]

Thanks, I'll have a look.

FYI: In the unit tests, you can access private methods - at least when the code is compiled with the Automake files, because -Dprivate=public is used. (I know this is ugly, but it allows better coverage without adjusting the tested code...)

[Teemperor]

Sounds good to me, updated the branch!

[coveralls]

Coverage remained the same when pulling a409ba9 on Teemperor:unicode_escaping into 0cd2ecf on nlohmann:master.

[Teemperor]

Just realized I effectively reverted your commit 0cd2ecf about the code style patch. I'll apply the changes to this patch.

[coveralls]

Coverage remained the same when pulling a866a9d on Teemperor:unicode_escaping into 0cd2ecf on nlohmann:master.

[nlohmann]

Thanks a lot! This Unicode business is really a pain!