better fault tolerance in typechecker for "not enough arguments" error

[scabug]

See: http://stackoverflow.com/questions/24851454/presentation-compiler-type-completion-in-method-call-with-more-than-one-argumen#comment38591097_24851454

And: retronym/scala@c1460f5

[scabug]

Imported From: https://issues.scala-lang.org/browse/SI-8739?orig=1
Reporter: @retronym

[scabug]

@adriaanm said:
We had a recent PR around arity mismatches by @som-snytt, I believe -- maybe that fixed this one too?

[scabug]

@som-snytt said (edited on Jun 23, 2016 7:00:17 AM UTC):
No, totally unrelated, but this is interesting. Still no completions in arg of f2. The WIP typechecks the available arg while erroring. Another idea is to insert the required commas. The IDE offers placeholders.

This effort on SO deserves preservation: "if i type comma after dot and then try again to complete after dot, it works!"

scala> def f(g: String => Int) = g("42")
f: (g: String => Int)Int

scala> f(x => x.
*  [snip...]                     

scala> def f2(g: String => Int, i: Int) = g("42") * i
f2: (g: String => Int, i: Int)Int

scala> f2(x => x.
!=   ##   +   ->   ==   asInstanceOf   ensuring   equals   formatted   getClass   hashCode   isInstanceOf   toString   ?

scala> f2(x => x.

scala> f2(x => x. , )
*   [etc...]

WIP repurposed scala/scala#5243

It drove me nuts when Eclipse couldn't autocomplete inside Java 8 lambdas.

[SethTisue]

scala/scala#5243 could perhaps be revived, if this is still an issue?