Fix #8431 (Make HKT GADT constraints actually work) #8522

abgruszecki · 2020-03-11T18:40:51Z

@smarter this one is ready to go! Going commit-by-commit will help. It turns out that we didn't actually need to care about SI-2712, since:

the problematic function is appOK
which is a nested function of canInstantiate
which is only entered with the RHS tycon, _if_the rhs tycon is a non-instantiated type variable

Fix gadt-uninjectivity test, as it wasn't clear what it was actually testing.

And a comment to boot.

abgruszecki added 2 commits March 22, 2020 18:11

Make HKT GADT constraints actually work

23844d3

Fix gadt-uninjectivity test, as it wasn't clear what it was actually testing.

Fix inter-dependencies in HK GADTs

f94606a

abgruszecki force-pushed the fix-#8431 branch from fa06287 to 31d7750 Compare March 23, 2020 08:11

One more test for new injectivity logic

146c453

And a comment to boot.

abgruszecki force-pushed the fix-#8431 branch from 31d7750 to 146c453 Compare March 23, 2020 10:43

abgruszecki marked this pull request as ready for review March 23, 2020 11:35

abgruszecki requested a review from smarter March 23, 2020 11:46

abgruszecki changed the title ~~WIP: Fix #8431~~ Fix #8431 (Make HKT GADT constraints actually work) Mar 23, 2020

OlivierBlanvillain assigned smarter Mar 26, 2020

smarter approved these changes Mar 26, 2020

View reviewed changes

smarter merged commit a324502 into scala:master Mar 26, 2020

smarter deleted the fix-#8431 branch March 26, 2020 18:02

Provide feedback