AoC: Surprising interpretation of a^b .f vs f a^b.

[fridis]

Stumbled over this during one of the Advent-Of-Code puzzles, but did not have time to create an issue immediately:

See this example

test_calltail is

  a := "a"
  b := "b"

  Any.f      => "({Any.this}.f)"
  Any.f(a)   => "({Any.this}.f $a)"
  f(x)       => "(f $x)"
  f(x,y)     => "(f $x $y)"

  String.prefix ^ => "(^{String.this})"
  String.infix ^(b String) => "({String.this}^$b)"

  say " 1. a^b.f   : {a^b.f}"
  say " 2. a^b .f  : {a^b .f}"
  say " 3. a ^ b.f : {a ^ b.f}"
  say " 4. a ^ b .f: {a ^ b .f}"
  say " 5. (a^b).f : {(a^b).f}"
  say " 6. (a^b) .f: {(a^b) .f}"
  say " 7. a.f^b   : {a.f^b}"
  say " 8. a.f ^ b : {a.f ^ b}"
  say " 9. a.f ^b  : {a.f ^b}"
  say "10. f a^b   : {f a^b}"
  say "11. f a ^ b : {f a ^ b}"
  say "12. f a ^b  : {f a ^b}"

Which produces quite intuitive evaluation order for most cases:

 > ./build/bin/fz test_calltail.fz
 1. a^b.f   : (a^(b.f))
 2. a^b .f  : (a^(b.f))
 3. a ^ b.f : (a^(b.f))
 4. a ^ b .f: (a^(b.f))
 5. (a^b).f : ((a^b).f)
 6. (a^b) .f: ((a^b).f)
 7. a.f^b   : ((a.f)^b)
 8. a.f ^ b : ((a.f)^b)
 9. a.f ^b  : (a.f (^b))
10. f a^b   : (f (a^b))
11. f a ^ b : ((f a)^b)
12. f a ^b  : (f a (^b))

The case that I found surprising is number 2., especially when compared to number 10. I think we should change the parser to parse a^b .f as (a^b).f.

Related to this, we might produce an error for code like 1+2 * 3 where the spacing disagrees with the operator precedence.