Convert t/pmc/float.t to PIR

[flh]

The attached patch rewrites t/pmc/float.t in PIR.

This version does not use "print" anymore to test for a negative zero: is the todo related to TT#313 still relevant?

[bacek]

Replying to flh:

The attached patch rewrites t/pmc/float.t in PIR. This version does not use "print" anymore to test for a negative zero: is the todo related to TT#313 still relevant?

Yes, it's still relevant. You can implement it using "todo" from parrot's Test::More

-- Bacek

[flh]

Patch updated. I had forgotten to remove a #TODO I had left for myself.

bacek: the test is todoed, I was just wondering whether this todo was still necessary or not.

By the way, the "divide by zero" test in the Perl version looked wrong, because it used something like:

$P1 = new ['Float']
$P1 = "0.0"

This doesn't create a Float PMC with value 0.0, but a String PMC with the string value "0.0" (scalar PMCs automagically morph to the right type when we set their value).

[dukeleto]

Applied as r41097, thanks flh++ !