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
[ConstraintSystem] Use '# of overloads' as a tie-breaker in ambiguities
The deepest expression is the one that introduced the ambiguity into
the chain, so depth and index should be deciding factors and number of
overloads - a tie-breaker, while choosing what to diagnose.
Resolves: rdar://94360230
varpi_s1=SpecialPi(pi_s) // expected-error {{argument passed to call that takes no arguments}}
587
587
588
-
varpi_f2=Float(getPi()) // expected-error {{ambiguous use of 'init(_:)'}}
589
-
varpi_d2=Double(getPi()) // expected-error {{ambiguous use of 'init(_:)'}}
588
+
varpi_f2=Float(getPi()) // expected-error {{ambiguous use of 'getPi()'}}
589
+
varpi_d2=Double(getPi()) // expected-error {{ambiguous use of 'getPi()'}}
590
590
varpi_s2:SpecialPi=getPi() // no-warning
591
591
592
592
varfloat=Float.self
593
-
varpi_f3= float.init(getPi()) // expected-error {{ambiguous use of 'init(_:)'}}
593
+
varpi_f3= float.init(getPi()) // expected-error {{ambiguous use of 'getPi()'}}
594
594
varpi_f4= float.init(pi_f)
595
595
596
596
vare=Empty(f) // expected-warning {{variable 'e' inferred to have type 'Empty', which is an enum with no cases}} expected-note {{add an explicit type annotation to silence this warning}} {{8-8=: Empty}}
0 commit comments