#3054. Add more test cases to closurization tests. Part 2. #3069
Conversation
| A a = A(); | ||
| final fa1 = a.m; | ||
| final fa2 = a.m; | ||
| Expect.notEquals(fa1, fa2); |
There was a problem hiding this comment.
Note that the spec wording says that the two tear-offs are equal if and only if they are identical. But it doesn't say that they can't be identical.
In this particular case we actually need to preserve the permission to optimize the extension methods: If an extension method doesn't actually depend on the receiver (as in this example: it just returns an argument and never uses this) then it must be OK for an implementation to compile this extension method to, for example, a simple top-level function like num compiled_m(int r1, {String p1 = ""}) => r1; and then return that (simple and constant) function object when the corresponding extension method is torn off.
In this case we'll have identical(fa1, fa2) (and hence also fa1 == fa2).
So we have to weaken the test to say that it's OK for fa1 and fa2 to be unequal, and it's also OK for them to be equal (and even identical, if anyone asks ;-).
As usual, evaluation of fa1 == fa2 shouldn't throw, so we shouldn't just skip the test entirely.
So perhaps this would do?:
| Expect.notEquals(fa1, fa2); | |
| Expect.isTrue(fa1 == fa2 || fa1 != fa2); |
There was a problem hiding this comment.
Thank you. Makes sense. I simply deleted this test. Expect.isTrue(fa1 == fa2 || fa1 != fa2); will hardly ever fail and will always be true. Therefore, this test does nothing except confirm that extension and extension type methods can be torn off. We have lots of other tests (including ones in this PR) that check this.
There was a problem hiding this comment.
Do we have a test that verifies that the operator == on these function objects isn't buggy and throws?
There was a problem hiding this comment.
Please clarify, how == may throw on such function objects?
There was a problem hiding this comment.
That would be a bug, but bugs is the main reason why we have tests.
| ET et = ET(0); | ||
| final fet1 = et.m; | ||
| final fet2 = et.m; | ||
| Expect.notEquals(fet1, fet2); |
...Extraction/Instance_Method_Closurization/method_closurization_positional_params_A03_t03.dart
Outdated
Show resolved
Hide resolved
...Extraction/Instance_Method_Closurization/method_closurization_positional_params_A03_t03.dart
Outdated
Show resolved
Hide resolved
...Extraction/Instance_Method_Closurization/method_closurization_positional_params_A03_t03.dart
Outdated
Show resolved
Hide resolved
| A a = A(); | ||
| final fa1 = a.m; | ||
| final fa2 = a.m; | ||
| Expect.notEquals(fa1, fa2); |
There was a problem hiding this comment.
Again, should allow this, but it should also allow optimization that causes fa1 == fa2.
| ET et = ET(0); | ||
| final fet1 = et.m; | ||
| final fet2 = et.m; | ||
| Expect.notEquals(fet1, fet2); |
| A a = A(); | ||
| final fa1 = a.m; | ||
| final fa2 = a.m; | ||
| Expect.notEquals(fa1, fa2); |
There was a problem hiding this comment.
Thank you. Makes sense. I simply deleted this test. Expect.isTrue(fa1 == fa2 || fa1 != fa2); will hardly ever fail and will always be true. Therefore, this test does nothing except confirm that extension and extension type methods can be torn off. We have lots of other tests (including ones in this PR) that check this.
...Extraction/Instance_Method_Closurization/method_closurization_positional_params_A03_t03.dart
Outdated
Show resolved
Hide resolved
2025-02-13 [email protected] dart-lang/co19#3057. Operator `==` tests renamed (dart-lang/co19#3075) 2025-02-13 [email protected] Fixes dart-lang/co19#3073. Add `asyncEnd()` after `cleanup` invocation in `asyncTest` (dart-lang/co19#3074) 2025-02-07 [email protected] dart-lang/co19#3054. Add more test cases to closurization tests. Part 2. (dart-lang/co19#3069) Cq-Include-Trybots: luci.dart.try:analyzer-linux-release-try Change-Id: Ic56a06a8f8ec1fc36b04b5250ad5a4cf698e5404 Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/409940 Reviewed-by: Alexander Thomas <[email protected]> Reviewed-by: Erik Ernst <[email protected]> Commit-Queue: Alexander Thomas <[email protected]>
No description provided.