No. Most of the edge cases arise from trying to parse an underspecified plaintext protocol. Everything in HTTP/2 is length-prefixed and unambiguously specified.

No. Most of the edge cases arise from trying to parse an underspecified plaintext protocol. Everything in HTTP/2 is length-prefixed and unambiguously specified. That makes it dramatically easier to write a compliant parser or client.