fix: Incompleteness in the resolver when detecting support for equality #6231
Conversation
| if (allowCommonSuperType) { | ||
| if (HaveCommonSuperPreType(aa, bb)) { | ||
| return true; | ||
| } | ||
| } else { | ||
| if (IsSuperPreTypeOf(aa, bb) || IsSuperPreTypeOf(bb, aa)) { | ||
| return true; | ||
| } |
There was a problem hiding this comment.
Another way to accomplish this is to leave the original comparison
if (IsSuperPreTypeOf(aa, bb) || IsSuperPreTypeOf(bb, aa)) {
return true;
}and then do
if (allowCommonSuperType && HaveCommonSuperPreType(aa, bb)) {
return true;
}where the initial comparison if (IsSuperPreTypeOf(t, u) || IsSuperPreTypeOf(u, t)) ... inside HaveCommonSuperPreType can be removed.
There was a problem hiding this comment.
Ah OK so now two types are comparable iff, additionally to the usual rules, they have a common supertype and allowCommonSuperType is true. And allowCommonSuperType is true everywhere except for conversions and type tests. But if I do a quick search through PreTypeResolve.Expressions.cs, I never see both allowConversion and allowCommonSuperType to be both true or false at the same time, so isn't one implied by the other? Or did I miss something?
|
You're right, we could do with just one boolean flag the way we use this method. I think of them as two independent relaxations of the basic rule, so I like exposing them as such, but I'm happy to merge the flags into one. |
3 participants
Fixes #6181.