Fixed binding issue where was binding to the in the let expression rat...

[jcollard]

...her than the parameter.

b was being shadowed by another binding and was causing the range to be broken. Renaming a and b to l and h fixes this.

[jcollard]

Also, fixed a range issue with floats.

[jcollard]

Also, fixed a range issue with floats.

[evancz]

Is it also plausible to rename b to base and keep a and b the same? I think that'd be my preferred route.

A lot of the names in this code were really crazy and I did not feel comfortable making them longer, or really touching them at all, because I did not have an intuition of what they meant.

[evancz]

Is it also plausible to rename b to base and keep a and b the same? I think that'd be my preferred route.

A lot of the names in this code were really crazy and I did not feel comfortable making them longer, or really touching them at all, because I did not have an intuition of what they meant.

[evancz]

Reasoning for a and b is that I want to express going in that these are numbers with no ordering assumed, and I feel like l and h are sort of non-standard number names and suggest an invariant that's not true.

[evancz]

Reasoning for a and b is that I want to express going in that these are numbers with no ordering assumed, and I feel like l and h are sort of non-standard number names and suggest an invariant that's not true.

[jcollard]

Yeah, changing b to base should work.

I actually didn't realize let was letrec not let* which lead to a lot of confusion debugging this problem. :-P

[jcollard]

Yeah, changing b to base should work.

I actually didn't realize let was letrec not let* which lead to a lot of confusion debugging this problem. :-P

[jcollard]

fwiw, I agree that a and b are better names.

[jcollard]

fwiw, I agree that a and b are better names.

[evancz]

Awesome, looks great! Thanks for sorting this out, and sorry for introducing these mistakes :/

[evancz]

Awesome, looks great! Thanks for sorting this out, and sorry for introducing these mistakes :/

[jcollard]

No problem. I am just glad we caught them before 0.14.

[jcollard]

No problem. I am just glad we caught them before 0.14.

[evancz]

One benefit of 0.14 is that we can release core independent of the compiler now. This one could have been 1.0.1 :)

[evancz]

One benefit of 0.14 is that we can release core independent of the compiler now. This one could have been 1.0.1 :)