Pangram: Add tests for accented characters and non-latin scripts

[isani]

On the csharp track, many pangram solutions use the built-in method IsLetter(char). This method matches more than a-z, which leads to wrong results when the input contains Unicode letters. There are currently no tests that contain letters outside the a-z range, and thus a range of bugs are not caught.

The following strings should be tested as not pangrams (because they don't contain an 'a'):

"äbcdefghijklmnopqrstuvwxyz"
"αbcdefghijklmnopqrstuvwxyz"

The following strings should be tested as pangrams (to ensure input with Unicode isn't just rejected):

"the quick brown fox jumps over the lazy dög"
"the quick brown fox jumps over the lazy dөg"

[NobbZ]

Due to very different support of unicode across languages the problem-specifications do only cover ASCII (we haven't agreed if we use traditional 7bit or enhenced 8bit yet ;), some people imply the one, some people imply the other when reading ASCII).

If a language does cope well with Unicode handling, then its in the responsibility of the tracks maintainers to avoid and test against unicode pitfalls of that language.

Can you therefore file a bug at the C# issue tracker?

cc @exercism/csharp

[coriolinus]

This has been discussed before in #428.

[NobbZ]

Oh, it seems as if I am allowed to close here, so I'll move the issue over to C# repo directly.

[NobbZ]

Issue moved to exercism/csharp #502 via ZenHub