alteration of the returned result of the EGL function mathlib.precision() (RTC 46267)

Environment: Windows
Product name: IBM? Rational? Business Developer
Program Version : 7.5.1.4
Description.
My EGL program code when running as generated Java with the function precision() returns unexpected results for us when called with parameter of type decimal but correct result if the parameter is numeric ie.

program tstedt type BasicProgram{}
Prec2 decimal(2,1);
Prec4 decimal(4,2);
Prec2n num(2,1);
Prec4n num(4,2);
function main()
// Decimal wrong
syslib.writeStdout("Precision should be 2: <" + mathLib.precision(Prec2) + ">");
syslib.writeStdout("Decimals should be 1: <" + mathLib.decimals
(Prec2) + ">");
syslib.writeStdout("Precision should be 4: <" + mathLib.
precision(Prec4) + ">");
syslib.writeStdout("Decimals should be 2: <" + mathLib.decimals
(Prec4) + ">");
// Numeric OK
syslib.writeStdout("Precision should be 2: <" + mathLib.
precision(Prec2n) + ">");
syslib.writeStdout("Decimals should be 1: <" + mathLib.decimals
(Prec2n) + ">");
syslib.writeStdout("Precision should be 4: <" + mathLib.
precision(Prec4n) + ">");
syslib.writeStdout("Decimals should be 2: <" + mathLib.decimals
(Prec4n) + ">");
end
end

Produces output on console:
Precision should be 2: <3>
Decimals should be 1: <1>
Precision should be 4: <5>
Decimals should be 2: <2>
Precision should be 2: <2>
Decimals should be 1: <1>
Precision should be 4: <4>
Decimals should be 2: <2>

RBD development have informed us (via PMR 71531,SGC,724) that this is working as designed ie.
A decimal variable stores its data this way, two digits are stored in each byte, except for the last byte which holds one digit and the sign. This means there are always an odd number of digits. When you declare a decimal with an even length in EGL, you actually get that length + 1, except the maximum length is 32.
Prec2 decimal(2,1) = 12.3; // This does NOT overflow. The variable is actually a decimal(3,1).
The money and pacf types use a similar storage format to decimal, so they work the same, Num and numc do not.