fix 13607 #13648
fix 13607 #13648
Conversation
dsyme
left a comment
dsyme
left a comment
There was a problem hiding this comment.
@vzarytovskii more ideas for tests noted
| [<InlineData("DateTime", "DateTime.Now")>] | ||
| [<InlineData("int", "1")>] | ||
| [<InlineData("Guid", "(Guid.NewGuid())")>] | ||
| [<InlineData("Byte", "0x1")>] |
There was a problem hiding this comment.
Byte-typed literals are 1uy. Curious this compiles and passes. I guess the object is an integer and the first type test fails but the second still succeeds.
There was a problem hiding this comment.
We should test this kind of logic when testing array types. Particularly signed/unsigned integer arrays which have strange rules in .net (int32[] can runtime-coerce to unit32[]) . Likewise arrays of enums. Also covariant coercions for reference array types.
There was a problem hiding this comment.
Byte-typed literals are 1uy. Curious this compiles and passes. I guess the object is an integer and the first type test fails but the second still succeeds.
Sigh, yeah, I fixed it locally but didn't push. Well, obj is bad :D
3 participants
Fixes #13607
The problem is if the type test on the struct succeeds, but the condition fails, then later type tests for related interfaces are assumed to fail.
This is actually noted as a case of concern in the comment but there was not a matching test for it, which shows that my testing for this change was inadequate - it should have at least covered every case identified in the comments.
Some baselines may need updating