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
The current substrait-validatorparse interface accepts a Buf object, but there is no equivalent function like validate that accepts an already decoded Plan object and validates it.
There is the parse_plan function, but (1) it's not public (it's pub in the parse module, but the parse module itself is not public), and (2) the external interface should probably be something a bit different, more like fn validate(plan: &Plan, config: &Config) -> ParseResult to match the public parse function.
Out of curiosity - is there any reason this isn't exposed? The one I can guess at is that it would expose the version of substrait that substrait-validator is depending on, and the caller would have to use the same one (or their Plan types wouldn't be the same type); that would limit the utility of this function, but still seems to me like it would be a nice complement to the parse function. Are there other reasons, or is this a good idea?
The text was updated successfully, but these errors were encountered:
The current
substrait-validator
parse
interface accepts aBuf
object, but there is no equivalent function likevalidate
that accepts an already decodedPlan
object and validates it.There is the
parse_plan
function, but (1) it's not public (it'spub
in theparse
module, but theparse
module itself is not public), and (2) the external interface should probably be something a bit different, more likefn validate(plan: &Plan, config: &Config) -> ParseResult
to match the publicparse
function.Out of curiosity - is there any reason this isn't exposed? The one I can guess at is that it would expose the version of
substrait
thatsubstrait-validator
is depending on, and the caller would have to use the same one (or theirPlan
types wouldn't be the same type); that would limit the utility of this function, but still seems to me like it would be a nice complement to theparse
function. Are there other reasons, or is this a good idea?The text was updated successfully, but these errors were encountered: