Merge pull request #7736 from cooldome/range_float_type

Language feature: range float types

tests/float/tfloatrange.nim

@@ -0,0 +1,49 @@
+discard """
+  cmd: "nim c -d:release --rangeChecks:on $file"
+  output: '''StrictPositiveRange
+float
+range fail expected
+range fail expected
+'''
+"""
+import math, fenv
+
+type
+  Positive = range[0.0..Inf]
+  StrictPositive = range[minimumPositiveValue(float)..Inf]
+  Negative32 = range[-maximumPositiveValue(float32) .. -1.0'f32]
+
+proc myoverload(x: float) =
+  echo "float"
+
+proc myoverload(x: Positive) =
+  echo "PositiveRange"
+
+proc myoverload(x: StrictPositive) =
+  echo "StrictPositiveRange"
+
+let x = 9.0.StrictPositive
+myoverload(x)
+myoverload(9.0)
+
+doAssert(sqrt(x) == 3.0)
+
+var z = -10.0
+try:
+  myoverload(StrictPositive(z))
+except:
+  echo "range fail expected"
+  
+  
+proc strictOnlyProc(x: StrictPositive): bool =
+  if x > 1.0: true else: false
+  
+let x2 = 5.0.Positive
+doAssert(strictOnlyProc(x2))
+
+try:
+  let x4 = 0.0.Positive
+  discard strictOnlyProc(x4)
+except:
+  echo "range fail expected"
+