... | ... | @@ -4,7 +4,9 @@ |
|
|
|
|
|
Those with an MR actually have code.
|
|
|
|
|
|
- #17357, #17248: Fix strictness of pattern synonyms and newtypes before thinking about refactoring the whole non-void constraint stuff
|
|
|
- #17248, #17376, !1976: Get rid of the special case for `-XEmptyCase`. Fix handling of newtypes because of testsuite failures. Some ground work for proper non-void constraint handling.
|
|
|
|
|
|
- #17357: Fix strictness of pattern synonyms. We came to the agreement that it's not worth the trouble and most useful pattern synonyms are strict anyway.
|
|
|
|
|
|
- Clean up `provideEvidenceForEquation`, define `ensureInhabited delta = null <$> provideEvidenceForEquation 1 delta`
|
|
|
- Apparently, `provideEvidenceForEquation` currently assumes that every COMPLETE set is inhabited, and thus implicitly assumes that `ensureInhabited` is true for that data type. So we can't actually just re-define on in terms of the other just yet.
|
... | ... | |