Fix regression in exhaustivity of HK types #18303
Conversation
dwijnand
force-pushed
the
space-engine-aux-hk
branch
from
30731c1 to
09df3ef
Compare
dwijnand
added
the
backport:nominated
There was a problem hiding this comment.
Dale and I discussed this over Zoom yesterday and today. The story has grown a bit long and complicated (original ticket, original PR, regression ticket, this PR) and there are multiple overlapping concerns (which is why a partial reversion is even an option).
I'm satisfied that this PR should be merged, since (to make a long story short) 1) it fixes the regression, 2) it's only reverting code that was added only quite recently anyway. So this seems mergeable and desirable to me, even if it's not necessairily the final chapter in this story.
So I'm hitting approve, but let's summon @liufengyun. Fengyun, you approved the original PR, do you want to take a look at this one as well?
@Kordyjan note the backport-nominated label here. This partially reverts a PR that was merged for 3.3.1, so it should at least be considered for 3.3.2 IMO.
| @@ -4,5 +4,7 @@ case class Fun[A, B](f: Exp[A => B]) extends Exp[A => B] | |||
| class Test { | |||
| def eval(e: Fun[Int, Int]) = e match { | |||
| case Fun(x: Fun[Int, Double]) => ??? // error | |||
| case Fun(x: Exp[Int => String]) => ??? // error | |||
| case _ => | |||
There was a problem hiding this comment.
Not sure why the 2nd case was removed before, restoring it seems to be more reasonable. Otherwise, it seems not easy to justify why the 1st case receives a warning.
There was a problem hiding this comment.
The warning on the first one became (and now revert away from) an unchecked warning, on the unchecked type test.
The motivating case (i16451) is complicated, because it involves unchecked type arguments. To fix the regression, I'm reverting the fix.
dwijnand
force-pushed
the
space-engine-aux-hk
branch
from
09df3ef to
0839123
Compare
As issues are assigned to release milestone, I'd rather just open a new issue - which I'll do after merging. |
SethTisue
changed the title
|
There is still a regression in 3.3.1-RC7 after this was backported: #18507 |
Kordyjan
removed
the
backport:nominated
Reverts a key part of #16958, which is a complicated case, to fix the regression(s).