Closed karenetheridge closed 4 years ago
We probably want unicode semantics everywhere (where \d
matches U+BENGALI DIGIT FOUR), which means we need to clearly document this fact, and the caveat to use [0-9]
or [a-z]
if ascii semantics are needed.
Released in 0.012.
actually no -- see https://github.com/json-schema-org/JSON-Schema-Test-Suite/pull/505
see https://github.com/json-schema-org/JSON-Schema-Test-Suite/pull/380 --
it looks like the expectation is that character classes should be ascii-only. This requires adding the /a or even /aa flag onto all regexes, so e.g. BENGALI DIGIT FOUR (U+09EA) is not matched by the pattern<-- verify all this!\d
.Test with and without the presence of the
unicode_strings
feature. Submit test cases back to the test suite if needed.