BUG: Inconsistent behavior with float and float32, float64

[rudolph9]

cue version 0227ac0

myFloatAndIntZero: float & 0
myFloatAndIntOne: float & 1
myFloatAndFloatZero: float & 0.0
myFloatAndFloatOne: float & 1.0
myFloat32AndIntZero: float32 & 0
myFloat32AndIntOne: float32 & 1
myFloat32AndFloatZero: float32 & 0.0
myFloat32AndFloatOne: float32 & 1.0
myFloat64AndIntZero: float64 & 0
myFloat64AndIntOne: float64 & 1
myFloat64AndFloatZero: float64 & 0.0
myFloat64AndFloatOne: float64 & 1.0

:!cue eval % --all --ignore
myFloatAndIntZero:     _|_ // conflicting values float and 0 (mismatched types float and int)
myFloatAndIntOne:      _|_ // conflicting values float and 1 (mismatched types float and int)
myFloatAndFloatZero:   0.0
myFloatAndFloatOne:    1.0
myFloat32AndIntZero:   0
myFloat32AndIntOne:    1
myFloat32AndFloatZero: 0.0
myFloat32AndFloatOne:  1.0
myFloat64AndIntZero:   0
myFloat64AndIntOne:    1
myFloat64AndFloatZero: 0.0
myFloat64AndFloatOne:  1.0

I haven't verified this but It appears that declaring something float32|float64 only makes the value number with a range constraint, not the more constrained value float with a range.

[rudolph9]

Opening new ticket that proposes unifying some more fundamental aspects of cue.

[cueckoo]

This issue has been migrated to cue-lang/cue#252.

For more details about CUE's migration to a new home, please see cue-lang/cue#1078.