[WIP] Suppress the strict comparison warning when comparing .. with a literal

[pepkin88]

Currently, a code like

1
  .. == 1
  1 == ..

will produce misleading warnings upon compilation:

WARNING: strict comparison of two different types will always be false: .. == 1
WARNING: strict comparison of two different types will always be false: 1 == ..

It does that because .. is interpreted as a callable literal, probably for compiler's simplicity. That's why I added a simple exception to the conditions when to show the warning. I also have checked for other symbols requiring attention, but I think it's only .. (I checked Literal::compile's switch cases and the Literal class in general).

I don't know how to test for warnings, to I left the commit without tests.

[vendethiel]

Most probably .. shouldn’t be a literal...

[vendethiel]

Would you mind going the extra mile, and instead not represent .. as a Literal?
There shouldn't be that much code to change (those are 3 links)

(I'm lying... There's more than this that relies on it being a Literal..)

[pepkin88]

I'll look into that.
For now, I'll leave it there. There's no rush for that fix :)

[vendethiel]

FWIW the issue is also visible with 1 == this.

[rhendric]

I suppose this ought not to be a Literal either?

[vendethiel]

Would fix a few other issues, yes.

[rhendric]

Oh, neat. Which ones did you have in mind?

[vendethiel]

~> `this`; too I think.

[rhendric]

(Assuming you meant double-backticks there, and GitHub ate them; if not, I'm very confused!)

Hmm, I don't see it. That this will end up inside a JS node, regardless of what kind of node a bare this is. To make that behave the same as ~> this, I think you'd have to change Fun::uses-this to parse the contents of JS children to see if they contain this. (Which doesn't sound worth doing to me, TBH.)

[vendethiel]

Oops, that's on me, I totally forgot JS solved those issues already. (and yes, I did not miss infix this)