The Inbox: Kernel-ar.528.mcz

Andreas Raab uploaded a new version of Kernel to project The Inbox:
http://source.squeak.org/inbox/Kernel-ar.528.mcz

==================== Summary ====================

Name: Kernel-ar.528
Author: ar
Time: 14 December 2010, 12:13:09.254 am
UUID: b083f6fc-3978-ce48-8cfd-5aa8634845de
Ancestors: Kernel-mtf.527

Experimental. Provide support for both signaling and quiet NaNs in various Float operations. Covered by the preference Float>>signalNaN.

=============== Diff against Kernel-mtf.527 ===============

Item was changed:
  Number variableWordSubclass: #Float
  instanceVariableNames: ''
+ classVariableNames: 'E Epsilon Halfpi Infinity Ln10 Ln2 MaxVal MaxValLn MinValLogBase2 NaN NegativeInfinity NegativeZero Pi RadiansPerDegree SignalNaN Sqrt2 ThreePi Twopi'
- classVariableNames: 'E Epsilon Halfpi Infinity Ln10 Ln2 MaxVal MaxValLn MinValLogBase2 NaN NegativeInfinity NegativeZero Pi RadiansPerDegree Sqrt2 ThreePi Twopi'
  poolDictionaries: ''
  category: 'Kernel-Numbers'!
 
  !Float commentStamp: '<historical>' prior: 0!
  My instances represent IEEE-754 floating-point double-precision numbers.  They have about 16 digits of accuracy and their range is between plus and minus 10^307. Some valid examples are:
 
  8.0 13.3 0.3 2.5e6 1.27e-30 1.27e-31 -12.987654e12
 
  Mainly: no embedded blanks, little e for tens power, and a digit on both sides of the decimal point.  It is actually possible to specify a radix for Squeak Float constants.  This is great for teaching about numbers, but may be confusing to the average reader:
 
  3r20.2 --> 6.66666666666667
  8r20.2 --> 16.25
 
  If you don't have access to the definition of IEEE-754, you can figure out what is going on by printing various simple values in Float hex.  It may help you to know that the basic format is...
  sign 1 bit
  exponent 11 bits with bias of 1023 (16r3FF) to produce an exponent
  in the range -1023 .. +1024
  - 16r000:
  significand = 0: Float zero
  significand ~= 0: Denormalized number (exp = -1024, no hidden '1' bit)
  - 16r7FF:
  significand = 0: Infinity
  significand ~= 0: Not A Number (NaN) representation
  mantissa 53 bits, but only 52 are stored (20 in the first word, 32 in the second).  This is because a normalized mantissa, by definition, has a 1 to the right of its floating point, and IEEE-754 omits this redundant bit to gain an extra bit of precision instead.  People talk about the mantissa without its leading one as the FRACTION, and with its leading 1 as the SIGNFICAND.
 
  The single-precision format is...
  sign 1 bit
  exponent 8 bits, with bias of 127, to represent -126 to +127
                      - 0x0 and 0xFF reserved for Float zero (mantissa is ignored)
                      - 16r7F reserved for Float underflow/overflow (mantissa is ignored)
  mantissa 24 bits, but only 23 are stored
  This format is used in FloatArray (qv), and much can be learned from the conversion routines, Float asIEEE32BitWord, and Float class fromIEEE32Bit:.
 
  Thanks to Rich Harmon for asking many questions and to Tim Olson, Bruce Cohen, Rick Zaccone and others for the answers that I have collected here.!

Item was changed:
  ----- Method: Float class>>initialize (in category 'class initialization') -----
  initialize
  "Float initialize"
  "Constants from Computer Approximations, pp. 182-183:
  Pi = 3.14159265358979323846264338327950288
  Pi/2 = 1.57079632679489661923132169163975144
  Pi*2 = 6.28318530717958647692528676655900576
  Pi/180 = 0.01745329251994329576923690768488612
  2.0 ln = 0.69314718055994530941723212145817657
  2.0 sqrt = 1.41421356237309504880168872420969808"
 
  Pi := 3.14159265358979323846264338327950288.
  Halfpi := Pi / 2.0.
  Twopi := Pi * 2.0.
  ThreePi := Pi * 3.0.
  RadiansPerDegree := Pi / 180.0.
 
  Ln2 := 0.69314718055994530941723212145817657.
  Ln10 := 10.0 ln.
  Sqrt2 := 1.41421356237309504880168872420969808.
  E := 2.718281828459045235360287471353.
 
  Epsilon := 0.000000000001.  "Defines precision of mathematical functions"
 
  MaxVal := 1.7976931348623157e308.
  MaxValLn := 709.782712893384.
  MinValLogBase2 := -1074.
 
  Infinity := MaxVal * MaxVal.
  NegativeInfinity := 0.0 - Infinity.
  NaN := Infinity - Infinity.
  NegativeZero := 1.0 / Infinity negated.
+
+ SignalNaN := true.
  !

Item was added:
+ ----- Method: Float class>>signalNaN (in category 'preferences') -----
+ signalNaN
+ <preference: 'Signal NaN'
+ category: 'general' "since there is no math/arithmetic category"
+ description: 'When enabled, generate NaNError when encountering NaN. When disabled, silently propagate NaN'
+ type: #Boolean>
+ ^SignalNaN!

Item was added:
+ ----- Method: Float class>>signalNaN: (in category 'preferences') -----
+ signalNaN: aBool
+ "Preference accessor"
+ SignalNaN := aBool.
+ !

Item was changed:
  ----- Method: Float>>arcCos (in category 'mathematical functions') -----
  arcCos
  "Answer the angle in radians."
  <primitive: 'primitiveArcCos' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  ^ Halfpi - self arcSin!

Item was changed:
  ----- Method: Float>>arcCosH (in category 'mathematical functions') -----
  arcCosH
  <primitive: 'primitiveArcCosH' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
+ ^self primitiveFailed
+ !
- ^self primitiveFailed!

Item was changed:
  ----- Method: Float>>arcSin (in category 'mathematical functions') -----
  arcSin
  "Answer the angle in radians."
  <primitive: 'primitiveArcSin' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  ((self < -1.0) or: [self > 1.0]) ifTrue: [self error: 'Value out of range'].
  ((self = -1.0) or: [self = 1.0])
  ifTrue: [^ Halfpi * self]
  ifFalse: [^ (self / (1.0 - (self * self)) sqrt) arcTan]!

Item was changed:
  ----- Method: Float>>arcSinH (in category 'mathematical functions') -----
  arcSinH
  <primitive: 'primitiveArcSinH' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
+ ^self primitiveFailed
+ !
- ^self primitiveFailed!

Item was changed:
  ----- Method: Float>>arcTan (in category 'mathematical functions') -----
  arcTan
  "Answer the angle in radians.
  Optional. See Object documentation whatIsAPrimitive."
  <primitive: 'primitiveArcTan' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
- self = self ifFalse:[^self error: 'arcTan is undefined for NaN'].
  self < 0.0 ifTrue: [ ^ 0.0 - (0.0 - self) arcTan ].
  ^self primitiveArcTan!

Item was changed:
  ----- Method: Float>>arcTan: (in category 'mathematical functions') -----
  arcTan: denominator
  "Answer the angle in radians.
  Optional. See Object documentation whatIsAPrimitive."
  | result |
  <primitive: 'primitiveArcTan2' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
+ denominator isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^denominator].
  (self = 0.0) ifTrue: [
  (denominator > 0.0)
  ifTrue: [ result := 0 ]
  ifFalse: [ result := Pi ]
  ] ifFalse: [
  (denominator = 0.0) ifTrue: [
  (self > 0.0)
  ifTrue: [ result := Halfpi ]
  ifFalse: [ result := Halfpi negated ]
  ] ifFalse: [
  (denominator > 0)
  ifTrue: [ result := (self / denominator) arcTan ]
  ifFalse: [ result := ((self / denominator) arcTan) + Pi ]
  ].
  ].
  ^ result.!

Item was changed:
  ----- Method: Float>>arcTanH (in category 'mathematical functions') -----
  arcTanH
  <primitive: 'primitiveArcTanH' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
+ ^self primitiveFailed
+ !
- ^self primitiveFailed!

Item was changed:
  ----- Method: Float>>cos (in category 'mathematical functions') -----
  cos
  "Answer the cosine of the receiver taken as an angle in radians."
  <primitive: 'primitiveCos' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
- self = self ifFalse:[^self error: 'cos is undefined for NaN'].
- self abs = Float infinity ifTrue:[^self error: 'cos is undefined for Infinity'].
  ^ (self + Halfpi) sin!

Item was changed:
  ----- Method: Float>>cosH (in category 'mathematical functions') -----
  cosH
  "Answer the cosine of the receiver taken as an angle in radians."
  <primitive: 'primitiveCosH' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
+ ^self primitiveFailed
+ !
- ^self primitiveFailed!

Item was changed:
  ----- Method: Float>>exp (in category 'mathematical functions') -----
  exp
  "Answer E raised to the receiver power.
  Optional. See Object documentation whatIsAPrimitive."
 
  <primitive: 'primitiveExp' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
- self = self ifFalse:[^self error: 'exp is undefined for NaN'].
  "For now, fall back to the Squeak version of exp if FloatMathPlugin is absent"
  ^self primitiveExp!

Item was changed:
  ----- Method: Float>>fractionPart (in category 'truncation and round off') -----
  fractionPart
  "Primitive. Answer a Float whose value is the difference between the
  receiver and the receiver's asInteger value. Optional. See Object
  documentation whatIsAPrimitive."
  <primitive: 'primitiveFractionalPart' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  ^self - self truncated asFloat!

Item was changed:
  ----- Method: Float>>hypot: (in category 'mathematical functions') -----
  hypot: arg
  "hypot(x,y) returns sqrt(x^2+y^2) with error less  than 1 ulps"
  <primitive: 'primitiveHypot' module: 'FloatMathPlugin'>
  arg isFloat ifFalse:[^self hypot: arg asFloat].
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
+ arg isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^arg].
+ ^self primitiveFailed
+ !
- ^self primitiveFailed!

Item was changed:
  ----- Method: Float>>ln (in category 'mathematical functions') -----
  ln
  "Answer the natural logarithm of the receiver.
  Optional. See Object documentation whatIsAPrimitive."
  <primitive: 'primitiveLogN' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  self <= 0.0 ifTrue: [^self error: 'ln is only defined for x > 0.0'].
  ^self primitiveLn!

Item was changed:
  ----- Method: Float>>log (in category 'mathematical functions') -----
  log
  "Answer the base 10 logarithm of the receiver."
  <primitive: 'primitiveLog10' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  ^ self ln / Ln10!

Item was changed:
  ----- Method: Float>>sin (in category 'mathematical functions') -----
  sin
  "Answer the sine of the receiver taken as an angle in radians.
  Optional. See Object documentation whatIsAPrimitive."
  <primitive: 'primitiveSin' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
- self = self ifFalse:[^self error: 'sin is undefined for NaN'].
  self abs = Float infinity ifTrue:[^self error: 'sin is undefined for Infinity'].
  ^self primitiveSin!

Item was changed:
  ----- Method: Float>>sinH (in category 'mathematical functions') -----
  sinH
  <primitive: 'primitiveSinH' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  ^self primitiveFailed!

Item was changed:
  ----- Method: Float>>tan (in category 'mathematical functions') -----
  tan
  "Answer the tangent of the receiver taken as an angle in radians."
  <primitive: 'primitiveTan' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
- self = self ifFalse:[^self error: 'tan is undefined for NaN'].
  ^ self sin / self cos!

Item was changed:
  ----- Method: Float>>tanH (in category 'mathematical functions') -----
  tanH
  <primitive: 'primitiveTanH' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  ^self primitiveFailed!

Item was changed:
  ----- Method: Float>>timesTwoPower: (in category 'mathematical functions') -----
  timesTwoPower: anInteger
  "Primitive. Answer with the receiver multiplied by 2.0 raised
  to the power of the argument.
  Optional. See Object documentation whatIsAPrimitive."
  <primitive: 'primitiveTimesTwoPower' module: 'FloatMathPlugin'>
+ self isNaN ifTrue:[SignalNaN ifTrue:[NaNError signal]. ^self].
  ^self primitiveTimesTwoPower: anInteger!