Decimal module and test cases were updated to their state of

the art.  It now complies latest specification and tests.

The only difference of this version with the one in the trunk
is that a small subset that hash tests were removed, because
they rely on modifications to core hash() function (see
issue 1182 for further details).

Lib/test/decimaltestdata/abs.decTest

 ------------------------------------------------------------------------
 -- abs.decTest -- decimal absolute value                              --
--- Copyright (c) IBM Corporation, 1981, 2004.  All rights reserved.   --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
 ------------------------------------------------------------------------
 -- Please see the document "General Decimal Arithmetic Testcases"     --
 -- at http://www2.hursley.ibm.com/decimal for the description of      --
@@ -17,7 +17,7 @@
 --   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
 --   mfc@uk.ibm.com                                                   --
 ------------------------------------------------------------------------
-version: 2.39
+version: 2.57
 
 -- This set of tests primarily tests the existence of the operator.
 -- Additon, subtraction, rounding, and more overflows are tested
@@ -106,9 +106,9 @@ absx214 abs  0.01E-999        ->   1E-1001   Subnormal
 absx215 abs  0.999E-999       ->   1.00E-999 Inexact Rounded Subnormal Underflow
 absx216 abs  0.099E-999       ->   1.0E-1000 Inexact Rounded Subnormal Underflow
 absx217 abs  0.009E-999       ->   1E-1001   Inexact Rounded Subnormal Underflow
-absx218 abs  0.001E-999       ->   0E-1001   Inexact Rounded Subnormal Underflow
-absx219 abs  0.0009E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow
-absx220 abs  0.0001E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow
+absx218 abs  0.001E-999       ->   0E-1001   Inexact Rounded Subnormal Underflow Clamped
+absx219 abs  0.0009E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow Clamped
+absx220 abs  0.0001E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow Clamped
 
 absx230 abs -1.00E-999        ->   1.00E-999
 absx231 abs -0.1E-999         ->   1E-1000   Subnormal
@@ -119,9 +119,9 @@ absx234 abs -0.01E-999        ->   1E-1001   Subnormal
 absx235 abs -0.999E-999       ->   1.00E-999 Inexact Rounded Subnormal Underflow
 absx236 abs -0.099E-999       ->   1.0E-1000 Inexact Rounded Subnormal Underflow
 absx237 abs -0.009E-999       ->   1E-1001   Inexact Rounded Subnormal Underflow
-absx238 abs -0.001E-999       ->   0E-1001   Inexact Rounded Subnormal Underflow
-absx239 abs -0.0009E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow
-absx240 abs -0.0001E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow
+absx238 abs -0.001E-999       ->   0E-1001   Inexact Rounded Subnormal Underflow Clamped
+absx239 abs -0.0009E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow Clamped
+absx240 abs -0.0001E-999      ->   0E-1001   Inexact Rounded Subnormal Underflow Clamped
 
 -- long operand tests
 maxexponent: 999

Lib/test/decimaltestdata/clamp.decTest

 ------------------------------------------------------------------------
 -- clamp.decTest -- clamped exponent tests (format-independent)       --
--- Copyright (c) IBM Corporation, 2000, 2003.  All rights reserved.   --
+-- Copyright (c) IBM Corporation, 2000, 2007.  All rights reserved.   --
 ------------------------------------------------------------------------
 -- Please see the document "General Decimal Arithmetic Testcases"     --
 -- at http://www2.hursley.ibm.com/decimal for the description of      --
@@ -17,7 +17,7 @@
 --   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
 --   mfc@uk.ibm.com                                                   --
 ------------------------------------------------------------------------
-version: 2.39
+version: 2.57
 
 -- This set of tests uses the same limits as the 8-byte concrete
 -- representation, but applies clamping without using format-specific
@@ -73,10 +73,10 @@ clam092 apply   1.1e-398                -> #0000000000000001  Subnormal Underflo
 clam093 apply   1.00000000001e-398      -> #0000000000000001  Subnormal Underflow Inexact Rounded
 clam094 apply   1.00000000000001e-398   -> #0000000000000001  Subnormal Underflow Inexact Rounded
 clam095 apply   1.000000000000001e-398  -> #0000000000000001  Subnormal Underflow Inexact Rounded
-clam096 apply   0.1e-398                -> #0000000000000000  Subnormal Underflow Inexact Rounded
-clam097 apply   0.00000000001e-398      -> #0000000000000000  Subnormal Underflow Inexact Rounded
-clam098 apply   0.00000000000001e-398   -> #0000000000000000  Subnormal Underflow Inexact Rounded
-clam099 apply   0.000000000000001e-398  -> #0000000000000000  Subnormal Underflow Inexact Rounded
+clam096 apply   0.1e-398                -> #0000000000000000  Subnormal Underflow Inexact Rounded Clamped
+clam097 apply   0.00000000001e-398      -> #0000000000000000  Subnormal Underflow Inexact Rounded Clamped
+clam098 apply   0.00000000000001e-398   -> #0000000000000000  Subnormal Underflow Inexact Rounded Clamped
+clam099 apply   0.000000000000001e-398  -> #0000000000000000  Subnormal Underflow Inexact Rounded Clamped
 
 -- Same again, negatives
 -- Nmax and similar
@@ -112,10 +112,10 @@ clam192 apply   -1.1e-398               -> #8000000000000001  Subnormal Underflo
 clam193 apply   -1.00000000001e-398     -> #8000000000000001  Subnormal Underflow Inexact Rounded
 clam194 apply   -1.00000000000001e-398  -> #8000000000000001  Subnormal Underflow Inexact Rounded
 clam195 apply   -1.000000000000001e-398 -> #8000000000000001  Subnormal Underflow Inexact Rounded
-clam196 apply   -0.1e-398               -> #8000000000000000  Subnormal Underflow Inexact Rounded
-clam197 apply   -0.00000000001e-398     -> #8000000000000000  Subnormal Underflow Inexact Rounded
-clam198 apply   -0.00000000000001e-398  -> #8000000000000000  Subnormal Underflow Inexact Rounded
-clam199 apply   -0.000000000000001e-398 -> #8000000000000000  Subnormal Underflow Inexact Rounded
+clam196 apply   -0.1e-398               -> #8000000000000000  Subnormal Underflow Inexact Rounded Clamped
+clam197 apply   -0.00000000001e-398     -> #8000000000000000  Subnormal Underflow Inexact Rounded Clamped
+clam198 apply   -0.00000000000001e-398  -> #8000000000000000  Subnormal Underflow Inexact Rounded Clamped
+clam199 apply   -0.000000000000001e-398 -> #8000000000000000  Subnormal Underflow Inexact Rounded Clamped
 
 -- zeros
 clam401 apply   0E-500                  -> 0E-398  Clamped
@@ -184,6 +184,20 @@ clam669 apply   9E+370                  -> 9.0E+370 Clamped
 clam671 apply   9E+369                  -> 9E+369
 clam673 apply   9E+368                  -> 9E+368
 
+-- subnormals clamped to 0-Etiny
+precision:   16
+maxExponent: 384
+minExponent: -383
+clam681 apply 7E-398     -> 7E-398 Subnormal
+clam682 apply 0E-398     -> 0E-398
+clam683 apply 7E-399     -> 1E-398 Subnormal Underflow Inexact Rounded
+clam684 apply 4E-399     -> 0E-398 Clamped Subnormal Underflow Inexact Rounded
+clam685 apply 7E-400     -> 0E-398 Clamped Subnormal Underflow Inexact Rounded
+clam686 apply 7E-401     -> 0E-398 Clamped Subnormal Underflow Inexact Rounded
+clam687 apply 0E-399     -> 0E-398 Clamped
+clam688 apply 0E-400     -> 0E-398 Clamped
+clam689 apply 0E-401     -> 0E-398 Clamped
+
 -- example from documentation
 precision:   7
 rounding:    half_even

Lib/test/decimaltestdata/class.decTest

+------------------------------------------------------------------------
+-- class.decTest -- Class operations                                  --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
+------------------------------------------------------------------------
+-- Please see the document "General Decimal Arithmetic Testcases"     --
+-- at http://www2.hursley.ibm.com/decimal for the description of      --
+-- these testcases.                                                   --
+--                                                                    --
+-- These testcases are experimental ('beta' versions), and they       --
+-- may contain errors.  They are offered on an as-is basis.  In       --
+-- particular, achieving the same results as the tests here is not    --
+-- a guarantee that an implementation complies with any Standard      --
+-- or specification.  The tests are not exhaustive.                   --
+--                                                                    --
+-- Please send comments, suggestions, and corrections to the author:  --
+--   Mike Cowlishaw, IBM Fellow                                       --
+--   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
+--   mfc@uk.ibm.com                                                   --
+------------------------------------------------------------------------
+version: 2.57
+
+-- [New 2006.11.27]
+
+precision:   9
+maxExponent: 999
+minExponent: -999
+extended:    1
+clamp:       1
+rounding:    half_even
+
+clasx001  class    0                        -> +Zero
+clasx002  class    0.00                     -> +Zero
+clasx003  class    0E+5                     -> +Zero
+clasx004  class    1E-1007                  -> +Subnormal
+clasx005  class  0.1E-999                   -> +Subnormal
+clasx006  class  0.99999999E-999            -> +Subnormal
+clasx007  class  1.00000000E-999            -> +Normal
+clasx008  class   1E-999                    -> +Normal
+clasx009  class   1E-100                    -> +Normal
+clasx010  class   1E-10                     -> +Normal
+clasx012  class   1E-1                      -> +Normal
+clasx013  class   1                         -> +Normal
+clasx014  class   2.50                      -> +Normal
+clasx015  class   100.100                   -> +Normal
+clasx016  class   1E+30                     -> +Normal
+clasx017  class   1E+999                    -> +Normal
+clasx018  class   9.99999999E+999           -> +Normal
+clasx019  class   Inf                       -> +Infinity
+
+clasx021  class   -0                        -> -Zero
+clasx022  class   -0.00                     -> -Zero
+clasx023  class   -0E+5                     -> -Zero
+clasx024  class   -1E-1007                  -> -Subnormal
+clasx025  class  -0.1E-999                  -> -Subnormal
+clasx026  class  -0.99999999E-999           -> -Subnormal
+clasx027  class  -1.00000000E-999           -> -Normal
+clasx028  class  -1E-999                    -> -Normal
+clasx029  class  -1E-100                    -> -Normal
+clasx030  class  -1E-10                     -> -Normal
+clasx032  class  -1E-1                      -> -Normal
+clasx033  class  -1                         -> -Normal
+clasx034  class  -2.50                      -> -Normal
+clasx035  class  -100.100                   -> -Normal
+clasx036  class  -1E+30                     -> -Normal
+clasx037  class  -1E+999                    -> -Normal
+clasx038  class  -9.99999999E+999           -> -Normal
+clasx039  class  -Inf                       -> -Infinity
+
+clasx041  class   NaN                       -> NaN
+clasx042  class  -NaN                       -> NaN
+clasx043  class  +NaN12345                  -> NaN
+clasx044  class   sNaN                      -> sNaN
+clasx045  class  -sNaN                      -> sNaN
+clasx046  class  +sNaN12345                 -> sNaN
+
+
+-- decimal64 bounds
+
+precision:   16
+maxExponent: 384
+minExponent: -383
+clamp:       1
+rounding:    half_even
+
+clasx201  class    0                        -> +Zero
+clasx202  class    0.00                     -> +Zero
+clasx203  class    0E+5                     -> +Zero
+clasx204  class    1E-396                   -> +Subnormal
+clasx205  class  0.1E-383                   -> +Subnormal
+clasx206  class  0.999999999999999E-383     -> +Subnormal
+clasx207  class  1.000000000000000E-383     -> +Normal
+clasx208  class   1E-383                    -> +Normal
+clasx209  class   1E-100                    -> +Normal
+clasx210  class   1E-10                     -> +Normal
+clasx212  class   1E-1                      -> +Normal
+clasx213  class   1                         -> +Normal
+clasx214  class   2.50                      -> +Normal
+clasx215  class   100.100                   -> +Normal
+clasx216  class   1E+30                     -> +Normal
+clasx217  class   1E+384                    -> +Normal
+clasx218  class   9.999999999999999E+384    -> +Normal
+clasx219  class   Inf                       -> +Infinity
+
+clasx221  class   -0                        -> -Zero
+clasx222  class   -0.00                     -> -Zero
+clasx223  class   -0E+5                     -> -Zero
+clasx224  class   -1E-396                   -> -Subnormal
+clasx225  class  -0.1E-383                  -> -Subnormal
+clasx226  class  -0.999999999999999E-383    -> -Subnormal
+clasx227  class  -1.000000000000000E-383    -> -Normal
+clasx228  class  -1E-383                    -> -Normal
+clasx229  class  -1E-100                    -> -Normal
+clasx230  class  -1E-10                     -> -Normal
+clasx232  class  -1E-1                      -> -Normal
+clasx233  class  -1                         -> -Normal
+clasx234  class  -2.50                      -> -Normal
+clasx235  class  -100.100                   -> -Normal
+clasx236  class  -1E+30                     -> -Normal
+clasx237  class  -1E+384                    -> -Normal
+clasx238  class  -9.999999999999999E+384    -> -Normal
+clasx239  class  -Inf                       -> -Infinity
+
+clasx241  class   NaN                       -> NaN
+clasx242  class  -NaN                       -> NaN
+clasx243  class  +NaN12345                  -> NaN
+clasx244  class   sNaN                      -> sNaN
+clasx245  class  -sNaN                      -> sNaN
+clasx246  class  +sNaN12345                 -> sNaN
+
+
+

Lib/test/decimaltestdata/compare.decTest

 ------------------------------------------------------------------------
--- compare.decTest -- decimal comparison                              --
--- Copyright (c) IBM Corporation, 1981, 2003.  All rights reserved.   --
+-- compare.decTest -- decimal comparison that allows quiet NaNs       --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
 ------------------------------------------------------------------------
 -- Please see the document "General Decimal Arithmetic Testcases"     --
 -- at http://www2.hursley.ibm.com/decimal for the description of      --
@@ -17,11 +17,11 @@
 --   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
 --   mfc@uk.ibm.com                                                   --
 ------------------------------------------------------------------------
-version: 2.39
+version: 2.57
 
 -- Note that we cannot assume add/subtract tests cover paths adequately,
 -- here, because the code might be quite different (comparison cannot
--- overflow or underflow, so actual subtractions are not necesary).
+-- overflow or underflow, so actual subtractions are not necessary).
 
 extended: 1
 
@@ -112,10 +112,10 @@ comx086 compare   2.0   2.0  -> 0
 -- now some cases which might overflow if subtract were used
 maxexponent: 999999999
 minexponent: -999999999
-comx090 compare  9.99999999E+999999999 9.99999999E+999999999  -> 0
-comx091 compare -9.99999999E+999999999 9.99999999E+999999999  -> -1
-comx092 compare  9.99999999E+999999999 -9.99999999E+999999999 -> 1
-comx093 compare -9.99999999E+999999999 -9.99999999E+999999999 -> 0
+comx095 compare  9.99999999E+999999999 9.99999999E+999999999  -> 0
+comx096 compare -9.99999999E+999999999 9.99999999E+999999999  -> -1
+comx097 compare  9.99999999E+999999999 -9.99999999E+999999999 -> 1
+comx098 compare -9.99999999E+999999999 -9.99999999E+999999999 -> 0
 
 -- some differing length/exponent cases
 comx100 compare   7.0    7.0    -> 0
@@ -265,6 +265,21 @@ comx448 compare   -8.0    -90E-1  -> 1
 comx449 compare   -8      -.9E+1  -> 1
 comx450 compare   -8      -90E-1  -> 1
 
+-- misalignment traps for little-endian
+comx451 compare      1.0       0.1  -> 1
+comx452 compare      0.1       1.0  -> -1
+comx453 compare     10.0       0.1  -> 1
+comx454 compare      0.1      10.0  -> -1
+comx455 compare      100       1.0  -> 1
+comx456 compare      1.0       100  -> -1
+comx457 compare     1000      10.0  -> 1
+comx458 compare     10.0      1000  -> -1
+comx459 compare    10000     100.0  -> 1
+comx460 compare    100.0     10000  -> -1
+comx461 compare   100000    1000.0  -> 1
+comx462 compare   1000.0    100000  -> -1
+comx463 compare  1000000   10000.0  -> 1
+comx464 compare  10000.0   1000000  -> -1
 
 -- testcases that subtract to lots of zeros at boundaries [pgr]
 precision: 40
@@ -362,7 +377,7 @@ comx568 compare    1E+12  1       ->  1
 comx569 compare    1E+13  1       ->  1
 comx570 compare    1E+14  1       ->  1
 comx571 compare    1E+15  1       ->  1
--- similar with an useful coefficient, one side only
+-- similar with a useful coefficient, one side only
 comx580 compare  0.000000987654321     1E-15    -> 1
 comx581 compare  0.000000987654321     1E-14    -> 1
 comx582 compare  0.000000987654321     1E-13    -> 1
@@ -712,6 +727,32 @@ comx906 compare  1e-777777777 -1e-411111111 ->  1
 comx907 compare -1e-777777777  1e-411111111 -> -1
 comx908 compare -1e-777777777 -1e-411111111 ->  1
 
+-- spread zeros
+comx910 compare   0E-383  0       ->  0
+comx911 compare   0E-383 -0       ->  0
+comx912 compare  -0E-383  0       ->  0
+comx913 compare  -0E-383 -0       ->  0
+comx914 compare   0E-383  0E+384  ->  0
+comx915 compare   0E-383 -0E+384  ->  0
+comx916 compare  -0E-383  0E+384  ->  0
+comx917 compare  -0E-383 -0E+384  ->  0
+comx918 compare   0       0E+384  ->  0
+comx919 compare   0      -0E+384  ->  0
+comx920 compare  -0       0E+384  ->  0
+comx921 compare  -0      -0E+384  ->  0
+comx930 compare   0E+384  0       ->  0
+comx931 compare   0E+384 -0       ->  0
+comx932 compare  -0E+384  0       ->  0
+comx933 compare  -0E+384 -0       ->  0
+comx934 compare   0E+384  0E-383  ->  0
+comx935 compare   0E+384 -0E-383  ->  0
+comx936 compare  -0E+384  0E-383  ->  0
+comx937 compare  -0E+384 -0E-383  ->  0
+comx938 compare   0       0E-383  ->  0
+comx939 compare   0      -0E-383  ->  0
+comx940 compare  -0       0E-383  ->  0
+comx941 compare  -0      -0E-383  ->  0
+
 -- Null tests
 comx990 compare 10  # -> NaN Invalid_operation
 comx991 compare  # 10 -> NaN Invalid_operation

Lib/test/decimaltestdata/copy.decTest

+------------------------------------------------------------------------
+-- copy.decTest -- quiet copy                                         --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
+------------------------------------------------------------------------
+-- Please see the document "General Decimal Arithmetic Testcases"     --
+-- at http://www2.hursley.ibm.com/decimal for the description of      --
+-- these testcases.                                                   --
+--                                                                    --
+-- These testcases are experimental ('beta' versions), and they       --
+-- may contain errors.  They are offered on an as-is basis.  In       --
+-- particular, achieving the same results as the tests here is not    --
+-- a guarantee that an implementation complies with any Standard      --
+-- or specification.  The tests are not exhaustive.                   --
+--                                                                    --
+-- Please send comments, suggestions, and corrections to the author:  --
+--   Mike Cowlishaw, IBM Fellow                                       --
+--   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
+--   mfc@uk.ibm.com                                                   --
+------------------------------------------------------------------------
+version: 2.57
+
+extended:    1
+precision:   9
+rounding:    half_up
+maxExponent: 999
+minExponent: -999
+
+-- Sanity check
+cpyx001 copy       +7.50  -> 7.50
+
+-- Infinities
+cpyx011 copy  Infinity    -> Infinity
+cpyx012 copy  -Infinity   -> -Infinity
+
+-- NaNs, 0 payload
+cpyx021 copy         NaN  -> NaN
+cpyx022 copy        -NaN  -> -NaN
+cpyx023 copy        sNaN  -> sNaN
+cpyx024 copy       -sNaN  -> -sNaN
+
+-- NaNs, non-0 payload
+cpyx031 copy       NaN10  -> NaN10
+cpyx032 copy      -NaN10  -> -NaN10
+cpyx033 copy      sNaN10  -> sNaN10
+cpyx034 copy     -sNaN10  -> -sNaN10
+cpyx035 copy       NaN7   -> NaN7
+cpyx036 copy      -NaN7   -> -NaN7
+cpyx037 copy      sNaN101 -> sNaN101
+cpyx038 copy     -sNaN101 -> -sNaN101
+
+-- finites
+cpyx101 copy          7   -> 7
+cpyx102 copy         -7   -> -7
+cpyx103 copy         75   -> 75
+cpyx104 copy        -75   -> -75
+cpyx105 copy       7.50   -> 7.50
+cpyx106 copy      -7.50   -> -7.50
+cpyx107 copy       7.500  -> 7.500
+cpyx108 copy      -7.500  -> -7.500
+
+-- zeros
+cpyx111 copy          0   -> 0
+cpyx112 copy         -0   -> -0
+cpyx113 copy       0E+4   -> 0E+4
+cpyx114 copy      -0E+4   -> -0E+4
+cpyx115 copy     0.0000   -> 0.0000
+cpyx116 copy    -0.0000   -> -0.0000
+cpyx117 copy      0E-141  -> 0E-141
+cpyx118 copy     -0E-141  -> -0E-141
+
+-- full coefficients, alternating bits
+cpyx121 copy   268268268        -> 268268268
+cpyx122 copy  -268268268        -> -268268268
+cpyx123 copy   134134134        -> 134134134
+cpyx124 copy  -134134134        -> -134134134
+
+-- Nmax, Nmin, Ntiny
+cpyx131 copy  9.99999999E+999   -> 9.99999999E+999
+cpyx132 copy  1E-999            -> 1E-999
+cpyx133 copy  1.00000000E-999   -> 1.00000000E-999
+cpyx134 copy  1E-1007           -> 1E-1007
+
+cpyx135 copy  -1E-1007          -> -1E-1007
+cpyx136 copy  -1.00000000E-999  -> -1.00000000E-999
+cpyx137 copy  -1E-999           -> -1E-999
+cpyx138 copy  -9.99999999E+999  -> -9.99999999E+999

Lib/test/decimaltestdata/copyabs.decTest

+------------------------------------------------------------------------
+-- copyAbs.decTest -- quiet copy and set sign to zero                 --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
+------------------------------------------------------------------------
+-- Please see the document "General Decimal Arithmetic Testcases"     --
+-- at http://www2.hursley.ibm.com/decimal for the description of      --
+-- these testcases.                                                   --
+--                                                                    --
+-- These testcases are experimental ('beta' versions), and they       --
+-- may contain errors.  They are offered on an as-is basis.  In       --
+-- particular, achieving the same results as the tests here is not    --
+-- a guarantee that an implementation complies with any Standard      --
+-- or specification.  The tests are not exhaustive.                   --
+--                                                                    --
+-- Please send comments, suggestions, and corrections to the author:  --
+--   Mike Cowlishaw, IBM Fellow                                       --
+--   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
+--   mfc@uk.ibm.com                                                   --
+------------------------------------------------------------------------
+version: 2.57
+
+extended:    1
+precision:   9
+rounding:    half_up
+maxExponent: 999
+minExponent: -999
+
+-- Sanity check
+cpax001 copyabs       +7.50  -> 7.50
+
+-- Infinities
+cpax011 copyabs  Infinity    -> Infinity
+cpax012 copyabs  -Infinity   -> Infinity
+
+-- NaNs, 0 payload
+cpax021 copyabs         NaN  -> NaN
+cpax022 copyabs        -NaN  -> NaN
+cpax023 copyabs        sNaN  -> sNaN
+cpax024 copyabs       -sNaN  -> sNaN
+
+-- NaNs, non-0 payload
+cpax031 copyabs       NaN10  -> NaN10
+cpax032 copyabs      -NaN15  -> NaN15
+cpax033 copyabs      sNaN15  -> sNaN15
+cpax034 copyabs     -sNaN10  -> sNaN10
+cpax035 copyabs       NaN7   -> NaN7
+cpax036 copyabs      -NaN7   -> NaN7
+cpax037 copyabs      sNaN101 -> sNaN101
+cpax038 copyabs     -sNaN101 -> sNaN101
+
+-- finites
+cpax101 copyabs          7   -> 7
+cpax102 copyabs         -7   -> 7
+cpax103 copyabs         75   -> 75
+cpax104 copyabs        -75   -> 75
+cpax105 copyabs       7.10   -> 7.10
+cpax106 copyabs      -7.10   -> 7.10
+cpax107 copyabs       7.500  -> 7.500
+cpax108 copyabs      -7.500  -> 7.500
+
+-- zeros
+cpax111 copyabs          0   -> 0
+cpax112 copyabs         -0   -> 0
+cpax113 copyabs       0E+6   -> 0E+6
+cpax114 copyabs      -0E+6   -> 0E+6
+cpax115 copyabs     0.0000   -> 0.0000
+cpax116 copyabs    -0.0000   -> 0.0000
+cpax117 copyabs      0E-141  -> 0E-141
+cpax118 copyabs     -0E-141  -> 0E-141
+
+-- full coefficients, alternating bits
+cpax121 copyabs   268268268        -> 268268268
+cpax122 copyabs  -268268268        -> 268268268
+cpax123 copyabs   134134134        -> 134134134
+cpax124 copyabs  -134134134        -> 134134134
+
+-- Nmax, Nmin, Ntiny
+cpax131 copyabs  9.99999999E+999   -> 9.99999999E+999
+cpax132 copyabs  1E-999            -> 1E-999
+cpax133 copyabs  1.00000000E-999   -> 1.00000000E-999
+cpax134 copyabs  1E-1007           -> 1E-1007
+
+cpax135 copyabs  -1E-1007          -> 1E-1007
+cpax136 copyabs  -1.00000000E-999  -> 1.00000000E-999
+cpax137 copyabs  -1E-999           -> 1E-999
+cpax199 copyabs  -9.99999999E+999  -> 9.99999999E+999

Lib/test/decimaltestdata/copynegate.decTest

+------------------------------------------------------------------------
+-- copyNegate.decTest -- quiet copy and negate                        --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
+------------------------------------------------------------------------
+-- Please see the document "General Decimal Arithmetic Testcases"     --
+-- at http://www2.hursley.ibm.com/decimal for the description of      --
+-- these testcases.                                                   --
+--                                                                    --
+-- These testcases are experimental ('beta' versions), and they       --
+-- may contain errors.  They are offered on an as-is basis.  In       --
+-- particular, achieving the same results as the tests here is not    --
+-- a guarantee that an implementation complies with any Standard      --
+-- or specification.  The tests are not exhaustive.                   --
+--                                                                    --
+-- Please send comments, suggestions, and corrections to the author:  --
+--   Mike Cowlishaw, IBM Fellow                                       --
+--   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
+--   mfc@uk.ibm.com                                                   --
+------------------------------------------------------------------------
+version: 2.57
+
+extended:    1
+precision:   9
+rounding:    half_up
+maxExponent: 999
+minExponent: -999
+
+-- Sanity check
+cpnx001 copynegate       +7.50  -> -7.50
+
+-- Infinities
+cpnx011 copynegate  Infinity    -> -Infinity
+cpnx012 copynegate  -Infinity   -> Infinity
+
+-- NaNs, 0 payload
+cpnx021 copynegate         NaN  -> -NaN
+cpnx022 copynegate        -NaN  -> NaN
+cpnx023 copynegate        sNaN  -> -sNaN
+cpnx024 copynegate       -sNaN  -> sNaN
+
+-- NaNs, non-0 payload
+cpnx031 copynegate       NaN13  -> -NaN13
+cpnx032 copynegate      -NaN13  -> NaN13
+cpnx033 copynegate      sNaN13  -> -sNaN13
+cpnx034 copynegate     -sNaN13  -> sNaN13
+cpnx035 copynegate       NaN70  -> -NaN70
+cpnx036 copynegate      -NaN70  -> NaN70
+cpnx037 copynegate      sNaN101 -> -sNaN101
+cpnx038 copynegate     -sNaN101 -> sNaN101
+
+-- finites
+cpnx101 copynegate          7   -> -7
+cpnx102 copynegate         -7   -> 7
+cpnx103 copynegate         75   -> -75
+cpnx104 copynegate        -75   -> 75
+cpnx105 copynegate       7.50   -> -7.50
+cpnx106 copynegate      -7.50   -> 7.50
+cpnx107 copynegate       7.500  -> -7.500
+cpnx108 copynegate      -7.500  -> 7.500
+
+-- zeros
+cpnx111 copynegate          0   -> -0
+cpnx112 copynegate         -0   -> 0
+cpnx113 copynegate       0E+4   -> -0E+4
+cpnx114 copynegate      -0E+4   -> 0E+4
+cpnx115 copynegate     0.0000   -> -0.0000
+cpnx116 copynegate    -0.0000   -> 0.0000
+cpnx117 copynegate      0E-141  -> -0E-141
+cpnx118 copynegate     -0E-141  -> 0E-141
+
+-- full coefficients, alternating bits
+cpnx121 copynegate  268268268         -> -268268268
+cpnx122 copynegate  -268268268        -> 268268268
+cpnx123 copynegate  134134134         -> -134134134
+cpnx124 copynegate  -134134134        -> 134134134
+
+-- Nmax, Nmin, Ntiny
+cpnx131 copynegate  9.99999999E+999   -> -9.99999999E+999
+cpnx132 copynegate  1E-999                     -> -1E-999
+cpnx133 copynegate  1.00000000E-999   -> -1.00000000E-999
+cpnx134 copynegate  1E-1007                    -> -1E-1007
+
+cpnx135 copynegate  -1E-1007                   -> 1E-1007
+cpnx136 copynegate  -1.00000000E-999  -> 1.00000000E-999
+cpnx137 copynegate  -1E-999                    -> 1E-999
+cpnx138 copynegate  -9.99999999E+999  -> 9.99999999E+999

Lib/test/decimaltestdata/copysign.decTest

+------------------------------------------------------------------------
+-- copysign.decTest -- quiet copy with sign from rhs                  --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
+------------------------------------------------------------------------
+-- Please see the document "General Decimal Arithmetic Testcases"     --
+-- at http://www2.hursley.ibm.com/decimal for the description of      --
+-- these testcases.                                                   --
+--                                                                    --
+-- These testcases are experimental ('beta' versions), and they       --
+-- may contain errors.  They are offered on an as-is basis.  In       --
+-- particular, achieving the same results as the tests here is not    --
+-- a guarantee that an implementation complies with any Standard      --
+-- or specification.  The tests are not exhaustive.                   --
+--                                                                    --
+-- Please send comments, suggestions, and corrections to the author:  --
+--   Mike Cowlishaw, IBM Fellow                                       --
+--   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
+--   mfc@uk.ibm.com                                                   --
+------------------------------------------------------------------------
+version: 2.57
+
+extended:    1
+precision:   9
+rounding:    half_up
+maxExponent: 999
+minExponent: -999
+
+-- Sanity check, and examples from decArith
+cpsx001 copysign   +7.50       11  -> 7.50
+cpsx002 copysign   '1.50'   '7.33' -> 1.50
+cpsx003 copysign  '-1.50'   '7.33' -> 1.50
+cpsx004 copysign   '1.50'  '-7.33' -> -1.50
+cpsx005 copysign  '-1.50'  '-7.33' -> -1.50
+
+-- Infinities
+cpsx011 copysign  Infinity       11 -> Infinity
+cpsx012 copysign  -Infinity      11 -> Infinity
+
+-- NaNs, 0 payload
+cpsx021 copysign         NaN     11 -> NaN
+cpsx022 copysign        -NaN     11 -> NaN
+cpsx023 copysign        sNaN     11 -> sNaN
+cpsx024 copysign       -sNaN     11 -> sNaN
+
+-- NaNs, non-0 payload
+cpsx031 copysign       NaN10     11 -> NaN10
+cpsx032 copysign      -NaN10     11 -> NaN10
+cpsx033 copysign      sNaN10     11 -> sNaN10
+cpsx034 copysign     -sNaN10     11 -> sNaN10
+cpsx035 copysign       NaN7      11 -> NaN7
+cpsx036 copysign      -NaN7      11 -> NaN7
+cpsx037 copysign      sNaN101    11 -> sNaN101
+cpsx038 copysign     -sNaN101    11 -> sNaN101
+
+-- finites
+cpsx101 copysign          7      11 -> 7
+cpsx102 copysign         -7      11 -> 7
+cpsx103 copysign         75      11 -> 75
+cpsx104 copysign        -75      11 -> 75
+cpsx105 copysign       7.50      11 -> 7.50
+cpsx106 copysign      -7.50      11 -> 7.50
+cpsx107 copysign       7.500     11 -> 7.500
+cpsx108 copysign      -7.500     11 -> 7.500
+
+-- zeros
+cpsx111 copysign          0      11 -> 0
+cpsx112 copysign         -0      11 -> 0
+cpsx113 copysign       0E+4      11 -> 0E+4
+cpsx114 copysign      -0E+4      11 -> 0E+4
+cpsx115 copysign     0.0000      11 -> 0.0000
+cpsx116 copysign    -0.0000      11 -> 0.0000
+cpsx117 copysign      0E-141     11 -> 0E-141
+cpsx118 copysign     -0E-141     11 -> 0E-141
+
+-- full coefficients, alternating bits
+cpsx121 copysign   268268268           11 -> 268268268
+cpsx122 copysign  -268268268           11 -> 268268268
+cpsx123 copysign   134134134           11 -> 134134134
+cpsx124 copysign  -134134134           11 -> 134134134
+
+-- Nmax, Nmin, Ntiny
+cpsx131 copysign  9.99999999E+999      11 -> 9.99999999E+999
+cpsx132 copysign  1E-999               11 -> 1E-999
+cpsx133 copysign  1.00000000E-999      11 -> 1.00000000E-999
+cpsx134 copysign  1E-1007              11 -> 1E-1007
+
+cpsx135 copysign  -1E-1007             11 -> 1E-1007
+cpsx136 copysign  -1.00000000E-999     11 -> 1.00000000E-999
+cpsx137 copysign  -1E-999              11 -> 1E-999
+cpsx138 copysign  -9.99999999E+999     11 -> 9.99999999E+999
+
+-- repeat with negative RHS
+
+-- Infinities
+cpsx211 copysign  Infinity       -34 -> -Infinity
+cpsx212 copysign  -Infinity      -34 -> -Infinity
+
+-- NaNs, 0 payload
+cpsx221 copysign         NaN     -34 -> -NaN
+cpsx222 copysign        -NaN     -34 -> -NaN
+cpsx223 copysign        sNaN     -34 -> -sNaN
+cpsx224 copysign       -sNaN     -34 -> -sNaN
+
+-- NaNs, non-0 payload
+cpsx231 copysign       NaN10     -34 -> -NaN10
+cpsx232 copysign      -NaN10     -34 -> -NaN10
+cpsx233 copysign      sNaN10     -34 -> -sNaN10
+cpsx234 copysign     -sNaN10     -34 -> -sNaN10
+cpsx235 copysign       NaN7      -34 -> -NaN7
+cpsx236 copysign      -NaN7      -34 -> -NaN7
+cpsx237 copysign      sNaN101    -34 -> -sNaN101
+cpsx238 copysign     -sNaN101    -34 -> -sNaN101
+
+-- finites
+cpsx301 copysign          7      -34 -> -7
+cpsx302 copysign         -7      -34 -> -7
+cpsx303 copysign         75      -34 -> -75
+cpsx304 copysign        -75      -34 -> -75
+cpsx305 copysign       7.50      -34 -> -7.50
+cpsx306 copysign      -7.50      -34 -> -7.50
+cpsx307 copysign       7.500     -34 -> -7.500
+cpsx308 copysign      -7.500     -34 -> -7.500
+
+-- zeros
+cpsx311 copysign          0      -34 -> -0
+cpsx312 copysign         -0      -34 -> -0
+cpsx313 copysign       0E+4      -34 -> -0E+4
+cpsx314 copysign      -0E+4      -34 -> -0E+4
+cpsx315 copysign     0.0000      -34 -> -0.0000
+cpsx316 copysign    -0.0000      -34 -> -0.0000
+cpsx317 copysign      0E-141     -34 -> -0E-141
+cpsx318 copysign     -0E-141     -34 -> -0E-141
+
+-- full coefficients, alternating bits
+cpsx321 copysign   268268268          -18 -> -268268268
+cpsx322 copysign  -268268268          -18 -> -268268268
+cpsx323 copysign   134134134          -18 -> -134134134
+cpsx324 copysign  -134134134          -18 -> -134134134
+
+-- Nmax, Nmin, Ntiny
+cpsx331 copysign  9.99999999E+999     -18 -> -9.99999999E+999
+cpsx332 copysign  1E-999              -18 -> -1E-999
+cpsx333 copysign  1.00000000E-999     -18 -> -1.00000000E-999
+cpsx334 copysign  1E-1007             -18 -> -1E-1007
+
+cpsx335 copysign  -1E-1007            -18 -> -1E-1007
+cpsx336 copysign  -1.00000000E-999    -18 -> -1.00000000E-999
+cpsx337 copysign  -1E-999             -18 -> -1E-999
+cpsx338 copysign  -9.99999999E+999    -18 -> -9.99999999E+999
+
+-- Other kinds of RHS
+cpsx401 copysign          701    -34 -> -701
+cpsx402 copysign         -720    -34 -> -720
+cpsx403 copysign          701    -0  -> -701
+cpsx404 copysign         -720    -0  -> -720
+cpsx405 copysign          701    +0  ->  701
+cpsx406 copysign         -720    +0  ->  720
+cpsx407 copysign          701    +34 ->  701
+cpsx408 copysign         -720    +34 ->  720
+
+cpsx413 copysign          701    -Inf  -> -701
+cpsx414 copysign         -720    -Inf  -> -720
+cpsx415 copysign          701    +Inf  ->  701
+cpsx416 copysign         -720    +Inf  ->  720
+
+cpsx420 copysign          701    -NaN  -> -701
+cpsx421 copysign         -720    -NaN  -> -720
+cpsx422 copysign          701    +NaN  ->  701
+cpsx423 copysign         -720    +NaN  ->  720
+cpsx425 copysign         -720    +NaN8 ->  720
+
+cpsx426 copysign          701    -sNaN  -> -701
+cpsx427 copysign         -720    -sNaN  -> -720
+cpsx428 copysign          701    +sNaN  ->  701
+cpsx429 copysign         -720    +sNaN  ->  720
+cpsx430 copysign         -720    +sNaN3 ->  720
+

Lib/test/decimaltestdata/decSingle.decTest

+------------------------------------------------------------------------
+-- decSingle.decTest -- run all decSingle decimal arithmetic tests    --
+-- Copyright (c) IBM Corporation, 1981, 2007.  All rights reserved.   --
+------------------------------------------------------------------------
+-- Please see the document "General Decimal Arithmetic Testcases"     --
+-- at http://www2.hursley.ibm.com/decimal for the description of      --
+-- these testcases.                                                   --
+--                                                                    --
+-- These testcases are experimental ('beta' versions), and they       --
+-- may contain errors.  They are offered on an as-is basis.  In       --
+-- particular, achieving the same results as the tests here is not    --
+-- a guarantee that an implementation complies with any Standard      --
+-- or specification.  The tests are not exhaustive.                   --
+--                                                                    --
+-- Please send comments, suggestions, and corrections to the author:  --
+--   Mike Cowlishaw, IBM Fellow                                       --
+--   IBM UK, PO Box 31, Birmingham Road, Warwick CV34 5JL, UK         --
+--   mfc@uk.ibm.com                                                   --
+------------------------------------------------------------------------
+version: 2.57
+
+-- decSingle tests
+dectest: dsBase
+dectest: dsEncode
+