After investigation I finally understand how to make "pass" 6.4.2 the way it's implemented in this test.
By itself the first frame cebae1bdb9cf83cebcceb5f4 are invalid / incomplet / badly cut UTF8 message. Because it's the first message ( opcode 1 ) and not the last (!fin). I wait for next message to confirm that the whole message concatenate from frame1 + frame2 are again not an UTF8. Now I can fail fast and pass the test.
Why first frame are accepted in testsuite?
For 6.4.3 and 6.4.4 could failed for the wrong reason because it's seem it didn't respect something else.
An unfragmented message consists of a single frame with the FIN
bit set (Section 5.2) and an opcode other than 0.
After investigation I finally understand how to make "pass" 6.4.2 the way it's implemented in this test.
By itself the first frame cebae1bdb9cf83cebcceb5f4 are invalid / incomplet / badly cut UTF8 message. Because it's the first message ( opcode 1 ) and not the last (!fin). I wait for next message to confirm that the whole message concatenate from frame1 + frame2 are again not an UTF8. Now I can fail fast and pass the test.
Why first frame are accepted in testsuite?
For 6.4.3 and 6.4.4 could failed for the wrong reason because it's seem it didn't respect something else.
Should those two test use opcode 1 instead of 0 ?