Lazy and friends swallow fatal errors during implicit resolution

[joroKr21]

c.inferImplicitValue seems to have quite an extreme opinion on what silent means:

trait Tc[A]
object Tc {
  implicit def materialize[A]: Tc[A] = macro materializeImpl[A]
  def materializeImpl[A](c: blackbox.Context): c.Tree =
    throw new OutOfMemoryError("Whoops, I ran out of memory during implicit resolution.")
}

object UseTc {
  implicitly[Tc[Int]] // out of memory
  implicitly[Lazy[Tc[Int]]] // missing implicit !?
}

Now what is the chance this ever happened? Very low.
Did it happen here: https://travis-ci.org/milessabin/shapeless/jobs/274500922? Who knows?

[joroKr21]

Or the problem is here: https://github.com/milessabin/shapeless/blob/master/core/src/main/scala/shapeless/lazy.scala#L320-L322

[milessabin]

Interesting! Yes, that, looks like it might be an explanation. If it is we should be able to make the 2.10.x problem go away by bumping up the heap and/or stacksize ... other things being equal I'd prefer to leave the Lazy's as they are in the example, at least for the time being.

[joroKr21]

#777 passed this time. Hopefully it continues to do so in the future. I don't give sbt more than 2g memory on my machine either.

Apart from that the issue with the swallowed error seems unfixable. silent = false just converts every error to "implicit search has failed. to find out the reason, turn on -Xlog-implicits". Very helpful... Maybe this can be fixed in scalac. But I can't find where the error is caught.

[joroKr21]

This is hit and miss. If you don't mind me asking, why do you want to keep the Lazy in tail position? It does nothing there.

[macalinao]

Thanks for reporting this. The solution for me was to add the following to my SBT_OPTS env var:

-Xmx6G -Xms6G -Xss2M

[joroKr21]

Wow, 6G is crazy 😱 I opened milessabin/kittens#61

[macalinao]

I think it was probably the stack size that fixed the issue. I agree 6G is overkill.

…

On Sat, Sep 30, 2017 at 22:33 Georgi Krastev ***@***.***> wrote: Wow, 6G is crazy 😱 I opened milessabin/kittens#61 <https://github.com/milessabin/kittens/issues/61> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#776 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAYfby_MVt0EgxQHCl9TgPuScGcdawUKks5snwgRgaJpZM4Po-HI> .

[joroKr21]

Fixed in scalac by scala/scala#6344, nothing to be done here