Github user paul-rogers commented on the issue:
https://github.com/apache/drill/pull/701
Thank you for the very clear explanation. It is now obvious that the
original lazy init design relied on the "found/not found" semantics of a simple
name match: no ambiguity there. However, there is ambiguity with the overload
rules. Clearly, at planning stage, we need to check that our entire list of
candidate functions is consistent with ZK. Your design for that seems quite
good.
I wonder, at execution time, do we save a fully qualified function name so
we can go back to the "found/not found" trick? That is, do we save the function
as "foo( INT, VARCHAR )" or do we repeat the type inference at runtime?
I will do a detailed review of the code next.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
