This repository was archived by the owner on Dec 14, 2018. It is now read-only.
This repository was archived by the owner on Dec 14, 2018. It is now read-only.
Consider InputFormatter behavior when we can't read the charset #3961
Closed
Description
An input formatter might be confused by a content type that is supported, with a charset
parameter that isn't supported (via SupportedEncodings
).
Today this results in the formatter returning true
from CanRead
and then adding a modelstate error when it's called to read the body.
We should just make this check part of CanRead
, then it would go down the normal 'unsupported content type' path when the encoding isn't supported.
We should additionally double-check the behavior where a formatter has an empty SupportedEncodings
. This should be allowed for binary content types that don't support/require a charset.