Inconsistent / broken pattern matching for unicode letters in PCRE mode

[rm-code]

Today I noticed, that RegExr seems to provide inconsistent / false matches for unicode pattern matching. Maybe I'm missing something...

1. Pattern with non-unicode string:

Pattern: [\p{L}]+ Test: Example

=> Match

2. Pattern with unicode string:

Pattern: [\p{L}]+ Test: Exämple

=> 2 Matches

3. Pattern with start + end tags and unicode string:

Pattern: ^[\p{L}]+$ Test: Exämple

=> No Match

4. Pattern with start tag and unicode string:

Pattern: ^[\p{L}]+ Test: Exämple

=> 1 Match