Issue #4084: Fix max, min, max_mag and min_mag Decimal methods to

give correct results in the case where one argument is a quiet NaN
and the other is a finite number that requires rounding.
Thanks Mark Dickinson.
This commit is contained in:
Facundo Batista 2008-12-11 04:20:07 +00:00
parent 10b24e87e5
commit 708d58179d
3 changed files with 37 additions and 16 deletions

View file

@ -154,6 +154,23 @@ extr1301 fma Inf 0 sNaN456 -> NaN Invalid_operation
extr1302 fma 0E123 -Inf sNaN789 -> NaN Invalid_operation
extr1302 fma -Inf 0E-456 sNaN148 -> NaN Invalid_operation
-- max/min/max_mag/min_mag bug in 2.5.2/2.6/3.0: max(NaN, finite) gave
-- incorrect answers when the finite number required rounding; similarly
-- for the other thre functions
maxexponent: 999
minexponent: -999
precision: 6
rounding: half_even
extr1400 max NaN 1234567 -> 1.23457E+6 Inexact Rounded
extr1401 max 3141590E-123 NaN1729 -> 3.14159E-117 Rounded
extr1402 max -7.654321 -NaN -> -7.65432 Inexact Rounded
extr1410 min -NaN -765432.1 -> -765432 Inexact Rounded
extr1411 min 3141592 NaN -> 3.14159E+6 Inexact Rounded
extr1420 max_mag 0.1111111 -NaN123 -> 0.111111 Inexact Rounded
extr1421 max_mag NaN999999999 0.001234567 -> 0.00123457 Inexact Rounded
extr1430 min_mag 9181716151 -NaN -> 9.18172E+9 Inexact Rounded
extr1431 min_mag NaN4 1.818180E100 -> 1.81818E+100 Rounded
-- Tests for the is_* boolean operations
precision: 9
maxExponent: 999