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 reasoning is that Test itself can never use the NonZero discriminant optimization itself, so it should use discriminants 1 and 2 to guarantee that it is itself never all-zero. This should be backwards-compatible as AFAIK safe code cannot observe, and unsafe code is not allowed to rely on, the concrete values of the discriminants.
(Not sure if an issue is enough for this or if it needs an RFC, so I started with the simpler-to-do one. ;)
The text was updated successfully, but these errors were encountered:
I would expect the following program to demonstrate that
Test
isNonZero
:The reasoning is that
Test
itself can never use theNonZero
discriminant optimization itself, so it should use discriminants 1 and 2 to guarantee that it is itself never all-zero. This should be backwards-compatible as AFAIK safe code cannot observe, and unsafe code is not allowed to rely on, the concrete values of the discriminants.(Not sure if an issue is enough for this or if it needs an RFC, so I started with the simpler-to-do one. ;)
The text was updated successfully, but these errors were encountered: