Underrun | UnsupportedHeaderValue |
NonCanonicalNumeric | InvalidSimpleValue | InvalidUtf8 |
NonCanonicalString | UnusedData | MisorderedMapKey |
DuplicateMapKey - see CborErrorDetailsByCode.
Decoded byte strings are zero-copy views aliasing the input
buffer - mutating the input after decoding (or mutating the returned
bytes) changes the other side. Call .slice() first if you need an
independent copy. This is deliberate: the zero-copy decode performance
profile is part of the library's contract.
Decode a single dCBOR item from
data, enforcing every deterministic encoding rule (canonical numeric forms, NFC text, map-key order, no trailing bytes). Throws CborError on any violation.