Update or remove non-ASCII test cases.

[Insti]

In #428 the decision was made to remove test cases containing non-ASCII characers from exercises that are not explicitly about extended character set handling.

The following tests were identified as containing test cases that need to be fixed:

• exercises/atbash-cipher/canonical-data.json
• exercises/bob/canonical-data.json
• exercises/forth/canonical-data.json
• exercises/isogram/canonical-data.json
• exercises/pangram/canonical-data.json
• exercises/run-length-encoding/canonical-data.json
• exercises/scrabble-score/canonical-data.json

Please check before removing a test that it is not also testing another property of the solution. If it is, please update the test to remove the non-ascii characters while maintaining the other intended purpose.

[ldwoolley]

The discussion is still open? isn't it? I think handling unicode where ever possible future proofs the solutions and improves the programming solutions.

[robkeim]

I just sent out #527 which addresses atbash-cipher, forth, and scrabble-score.

The last exercise isogram was already addressed in #433.