Implement floating-point addition and subtraction without using double type.

[brazilofmux]

Original issue 205 created by brazilofmux on 2006-10-30T23:10:59.000Z:

Because IEEE format is base-2, subtraction would be more accurate if
performed using base-10 operations instead of converting to and from a
double type. It would need to implement all the precision, range, and
special names of the existing 'double' code.

[brazilofmux]

Comment #1 originally posted by brazilofmux on 2006-12-27T20:51:11.000Z:

<empty>

[brazilofmux]

Comment #2 originally posted by brazilofmux on 2007-09-21T01:01:19.000Z:

Not critical for 2.7.

[brazilofmux]

Comment #3 originally posted by brazilofmux on 2010-09-25T21:16:28.000Z:

Issue 645 has been merged into this issue.