gdb/866: doublest.c should be replaced with a real floating point library

[drow at mvista dot com]

>Number:         866
>Category:       gdb
>Synopsis:       doublest.c should be replaced with a real floating point library
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Dec 03 21:38:00 PST 2002
>Closed-Date:
>Last-Modified:
>Originator:     drow@mvista.com
>Release:        5.3
>Organization:
>Environment:

>Description:
The method we use to construct doubles assumes IEEE behavior
in some places: for instance it adds a normalized double
to a denormal, which can trap on Alpha in non-IEEE mode.

There's a library in GCC (real.c) which we could be using for
this and for all our other floatformat needs.  It would be
nice to integrate them.
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: