moduleClassOf[O.type]

[Jason Zaugg]

scala> object O

defined module O

scala> def impl[T: c.WeakTypeTag](c: Context): c.Expr[Class[_]] = {import c.universe._; c.Expr[Class[_]](Literal(Constant(weakTypeOf[T])))}

impl: [T](c: scala.reflect.macros.Context)(implicit evidence$1: c.WeakTypeTag[T])c.Expr[Class[_]]

scala> def moduleClassOf[T <: Singleton]: Class[_] = macro impl[T]

moduleClassOf: [T <: Singleton]=> Class[_]

scala> moduleClassOf[O.type]

res5: Class[O.type] = class O$

I guess that it's something of an accident that this works, but I need to use this trick to fix SI-6259 in a manner more pleasing the the byte code verification police.

This also provides a tenable solution for SI-2453, where I just noticed my ponderings for last year:

> I wonder, could classOf be rewritten these days as a macro? 

> How about instanceClassOf / staticClassOf?

 

Hey, past me: it seems like `instanceClassOf` can!

-jason

[1]https://issues.scala-lang.org/browse/SI-2453?focusedCommentId=59963&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-59963

[Paul Phillips]

On Thu, Jan 24, 2013 at 12:30 PM, Jason Zaugg <jza...@gmail.com> wrote:

Hey, past me: it seems like `instanceClassOf` can!

Hey, present me: following up on those comments of yours is what has hung up the return of my "classOf as a macro" pr. But where it also led me is back to this:

  https://groups.google.com/d/msg/scala-language/ImqJuGylyi8/P4JGcNY1I_4J

If you listen carefully, you can still hear "past me":

  "It would be VERY USEFUL if there was! was... was... was..."