p4lang
Specify core.p4 match_kind semantics
@jafingerhut AI for Andy: check whether anywhere in the spec already says that when a table has multiple keys, an apply() operation only matches an entry if ALL the individual field matching criteria are true. That is, logically there is an AND between the matching conditions between multiple keys of a table.
Can we get some more comments on this PR? We would like to merge this in the next design meeting. I will apply the suggested changes.
@jafingerhut @vgurevich Pinging us to remind us to review the latest version of this PR, last updated on 2022-Nov-04.
@jafingerhut we are supposed to merge this but we need an approval from you
I did add a couple of review comments on 2022-Dec-05, and at least one of them I consider serious enough NOT to approve this PR in its current form (the one about the implied priority-per-field, which is untrue).
I would prefer if changes could be made to this PR to resolve those comments before I approve, but if others want to approve this as it is, I won't stop them.
Sorry, lost track which comments have been addressed. Any others I should be heeding?
@jafingerhut hopefully this is better. I had some conflicts when rebasing, hopefully I didn't break anything.