Re: [scala-debate] Extremely slow patmat phase of compilation

[Jon Pretty]

It looks like exhaustivity checking: removing `sealed` from the trait allows me to compile it in 43s. As I'm typing this, the `sealed` version is still compiling after five minutes.

Here's a pretty graph of how long it took to compile different numbers of cases with the supertrait `sealed` and not.

That doesn't look linear to me. The culprit should be relatively easy to find with a profiler...

Cheers,
Jon

On Tue Feb 24 2015 at 21:50:43 Rado Buranský <radobu...@gmail.com> wrote:

Can anybody tell me why this code takes endless time  to compile? What is the compiler doing so much and how can I help it?

You can find full source code here:

https://gist.github.com/RadoBuransky/d1d763a8f3c6555ee89a

sealed trait C
case object C1 extends C
case object C2 extends C
case object C3 extends C

...

case object C399 extends C
case object C400 extends C

object M {
  def f(c: C): Int = c match {
    case C1 => 1
    case C2 => 2
    case C3 => 3
    ...
    case C399 => 399
    case C400 => 400
  }
}

--
You received this message because you are subscribed to the Google Groups "scala-debate" group.
To unsubscribe from this group and stop receiving emails from it, send an email to scala-debate+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jon Pretty]

I've just been notified on Twitter by a friendly observer that this doesn't happen in Scala 2.11.4, so it's a recent regression.

You can avoid this by downgrading to Scala 2.11.4 or using -Xno-patmat-analysis (which turns off exhaustivity checking).

I've filed an issue here: https://issues.scala-lang.org/browse/SI-9181

Cheers,
Jon

[Jason Zaugg]

On Wed, Feb 25, 2015 at 8:01 AM, Jon Pretty <prope...@gmail.com> wrote:

I've just been notified on Twitter by a friendly observer that this doesn't happen in Scala 2.11.4, so it's a recent regression.

You can avoid this by downgrading to Scala 2.11.4 or using -Xno-patmat-analysis (which turns off exhaustivity checking).

Or more locally by (scrutinee: Any) match { ... }.

​

I've filed an issue here: https://issues.scala-lang.org/browse/SI-9181

Thanks, Jon.

-jason

[Rado Buranský]

Thanks a lot