Add some diagnostics related to lifetime dependence and function types #74956
Conversation
meg-gupta
requested review from
hborla,
slavapestov,
xedin and
AnthonyLatsis
as code owners
|
@swift-ci smoke test |
meg-gupta
force-pushed
the
lifetimedepfunctiontype
branch
from
0a4dd8c to
13cfbca
Compare
|
@swift-ci test |
| @@ -3248,6 +3248,10 @@ ConstraintSystem::matchFunctionTypes(FunctionType *func1, FunctionType *func2, | |||
| if (!matchFunctionIsolations(func1, func2, kind, flags, locator)) | |||
| return getTypeMatchFailure(locator); | |||
|
|
|||
| if (func1->getLifetimeDependenceInfo() != func2->getLifetimeDependenceInfo()) { | |||
| return getTypeMatchFailure(locator); | |||
There was a problem hiding this comment.
I think this works because the solver would use a "catch all" conversion fix for arguments but maybe we should add a tailored fix for this to indicate what really mismatched here instead of making developers compare types visually and guess?
There was a problem hiding this comment.
Are you suggesting to use a ContextualFailure? I didn't see precedent in CSSimplify where a ContextualFailure was used without attempting a ConstraintFix. Most type mismatches without ConstraintFix were flagged with the generic getTypeMatchFailure .
If it is okay to use a ContextualFailure here, I can do that as a follow on PR.
There was a problem hiding this comment.
I didn't really want to implement a ConstraintFix here. It can cause cascaded SIL diagnostics later on and maybe even more confusing to the user.
rdar://126169735