Even this I-D presents some interesting potential compatibility issues when round-tripping from canonical JSON to an internal representation and back, such as an implementation normalizing external text, or storing numbers in numerical types other than double.
The bigger issue in the past has not been the body of work defining a canonicalization scheme, but that people did that in the past for XML signatures. It became a huge interoperability issue due not just to the complexity of the canonicalization scheme[s], but to the difficulty in detecting and dealing with modifications to the canonical form. This also gracefully went down the slipper slope to signing portions and even specific features of a document, and tooling needing to accommodate preserving canonical form when e.g. signed documents were placed inside other documents. -DW > On Oct 10, 2018, at 6:29 PM, Manger, James <[email protected]> > wrote: > > https://tools.ietf.org/html/draft-rundgren-json-canonicalization-scheme > <https://tools.ietf.org/html/draft-rundgren-json-canonicalization-scheme>
_______________________________________________ jose mailing list [email protected] https://www.ietf.org/mailman/listinfo/jose
