non / imp Goto Github PK

This very small library has all the hallmarks of one that will be used by many, many libraries. So please can we have a Scala.js version sooner rather than later, please?

Hi there,

I just found out about this project and I find it very interesting; mostly because I didn't even know there was a runtime cost for using implicitly !

Do you have any benchmarks that you can share comparing imp with implicitly?

Thanks!

Currently the tests assume that they are run from the top-level of the repo.

However, it seems like during the release command the tests run with a different working directory (which means they fail). For now I have disabled the tests during releasing, but it would be better to figure out how to make it work correctly.

This looks like an oversight, since:

• previous versions have artifacts published for 2.10
• cross building appears to be set up

While adding scala.js support, I was forced to remove the src/main/scala_2.10 and src/main/scala_2.11 directories. Previously we used:

unmanagedSourceDirectories in Compile +=
  (sourceDirectory in Compile).value / s"scala_${scalaBinaryVersion.value}"

This didn't work with scala.js, so I was forced to write macros that worked in both 2.10 and 2.11. However, this makes the macros under 2.11 slightly less powerful.

Old code (more powerful):

def summon[Ev](c: Context)(ev: c.Tree): c.Tree = ev

New code (2.10 compatible):

def summon[Ev](c: Context)(ev: c.Expr[Ev]): c.Expr[Ev] = ev

Stephen Compall suggests that instead of returning Ev that the imp method should return Ev with ev.type to avoid losing refinements.