Initial rough support for deducing generic arguments in a call to a generic function. #4184
Conversation
generic function.
There was a problem hiding this comment.
FYI I haven't reviewed the generated SemIR yet; I'm looking for guidance about how best to do that.
Also FYI, I'm on vacation next week, so if we don't finish this today you may need to find another reviewer.
| // CHECK:STDOUT: %U.ref.loc15: type = name_ref U, %U.loc15_10.2 [symbolic = @Get.%U.1 (constants.%U)] | ||
| // CHECK:STDOUT: %.loc15_28.1: %.3 = tuple_literal (%T.ref.loc15, %U.ref.loc15) | ||
| // CHECK:STDOUT: %.loc15_28.2: type = converted %.loc15_28.1, constants.%.4 [symbolic = @Get.%.1 (constants.%.4)] | ||
| // CHECK:STDOUT: %return.var.loc15: ref @Get.%.1 (%.4) = var <return slot> |
There was a problem hiding this comment.
Should this %.4 be constants.%.4? If not, I can't find what it's referring to.
There was a problem hiding this comment.
Names in SemIR formatted types are by default in the constants scope, because type constants are almost always in that scope. (This can be confusing, but it does substantially reduce the verbosity. Maybe we should reconsider, perhaps using a shorter name for constants?)
There was a problem hiding this comment.
I don't have a strong opinion about how to spell names that refer to constants, but what really threw me off here was the fact that the previous line uses the longer form constants.%.4. Is there some significance to that? If not, it would be nice to fix the inconsistency one way or the other, but presumably that's out of scope for this PR.
| // CHECK:STDOUT: %U.ref.loc15: type = name_ref U, %U.loc15_10.2 [symbolic = @Get.%U.1 (constants.%U)] | ||
| // CHECK:STDOUT: %.loc15_28.1: %.3 = tuple_literal (%T.ref.loc15, %U.ref.loc15) | ||
| // CHECK:STDOUT: %.loc15_28.2: type = converted %.loc15_28.1, constants.%.4 [symbolic = @Get.%.1 (constants.%.4)] | ||
| // CHECK:STDOUT: %return.var.loc15: ref @Get.%.1 (%.4) = var <return slot> |
There was a problem hiding this comment.
I don't have a strong opinion about how to spell names that refer to constants, but what really threw me off here was the fact that the previous line uses the longer form constants.%.4. Is there some significance to that? If not, it would be nice to fix the inconsistency one way or the other, but presumably that's out of scope for this PR.
No description provided.