RESP Protocol Parser Practice

Return an error, because the bulk string `na` is shorter than its declared four bytes, making the frame malformed.

Return an error, because a RESP array is only valid when it arrives within a single TCP segment.

Wait, because a parser must always see a trailing newline before judging any frame, whatever the declared lengths say.

It should wait for more bytes because the declared bulk length says the payload has not arrived completely yet.

Splitting on spaces is fine for inline commands, and the only real problem is its performance on long strings whose argument scanning cost grows with the input length.

Splitting on spaces loses RESP framing and quoted or binary-safe payload boundaries; the parser must respect explicit lengths.

The space split fails only because the TCP layer strips quote characters from the byte stream in transit.

Splitting works if clients escape their spaces, since RESP's length prefixes exist only for binary payloads.

Consumed should cover only the complete arrays actually emitted, leaving the half third frame in the buffer for a later read.

Consumed should include the half frame too, since the parser already examined those bytes while scanning.

Consumed should be zero until the buffer holds nothing but complete frames, keeping the bookkeeping simple.

Consumed should always equal the total bytes read from the socket, with the parser tracking its own position separately.

The RESP parser should reject it, since the grammar of commands forbids arrays inside arrays, and rejecting early spares dispatch from handling malformed shapes.

Neither layer should reject it; the receiving handler can coerce nested arrays into flat strings as needed.

Command dispatch should reject a nested array as an invalid command argument after the RESP parser has faithfully decoded the RESP value.

The parser should silently flatten nested arrays, so that dispatch only ever sees strings arrive.

Both encode as $0, and clients tell them apart by a flag byte that follows the length line.

A null bulk string is encoded as $-1, while an empty bulk string is encoded with length zero followed by an empty payload.

A null bulk string is the absence of any reply at all, while an empty one is sent as length zero.

Null is sent as the literal word nil in a simple string, and empty as a simple string with no characters, matching what interactive clients print for missing keys.