Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | more tests for SI-3761 | Lukas Rytz | 2012-05-23 | 1 | -1/+17 |
| | |||||
* | SI-3761: Overload resolution fails on by-name parameter (amended per lrytz) | Som Snytt | 2012-05-22 | 1 | -0/+5 |
| | | | | Lukas noted the stopgap at methTypeArgs isn't needed. What! No black tie formals. | ||||
* | SI-3761: Overload resolution fails on by-name parameter | Som Snytt | 2012-05-21 | 1 | -0/+18 |
When isAsSpecific checks if method m applies to args of types of formal params of m1, a by-name parameter was converted to its underlying result type for the params (of m) but not the args (of m1). This had the useful effect of making m(A) more specific than m(=>A), which is the specified prioritization for implicit views, but also made m(=>A) and m(=>A, B*) ambiguous. To handle this edge case, the isCompatible test for A and =>A is made explicit, and by-name params are no longer converted. |