You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, there is no clean way to test that a refactoring of represented types does not change the binary encoding of those types. This requires us to be very cautious when changing types that might effect production systems – e.g. Tezos – and is particularly problematic when using ppx_repr, where the type representation does not show up in diffs.
To fix this problem, we should expose something like a "witness" of a particular binary codec that can be checked for equality:
with the idea being that equality of shapes implies equivalence of derived binary codecs. User libraries could then write expect tests of the digests of these shapes, providing some assurance of codec compatibility across revisions.
The text was updated successfully, but these errors were encountered:
Currently, there is no clean way to test that a refactoring of represented types does not change the binary encoding of those types. This requires us to be very cautious when changing types that might effect production systems – e.g. Tezos – and is particularly problematic when using
ppx_repr
, where the type representation does not show up in diffs.To fix this problem, we should expose something like a "witness" of a particular binary codec that can be checked for equality:
with the idea being that equality of shapes implies equivalence of derived binary codecs. User libraries could then write expect tests of the digests of these
shape
s, providing some assurance of codec compatibility across revisions.The text was updated successfully, but these errors were encountered: