Floating point trouble with x86's extended precision

[Gabriel Dos Reis]

dewar@gnat.com (Robert Dewar) writes:

| > It is OK to say, "Oh yes, this is a functionality we don't know how to
| > get right; we don't foresee any immediate work to fix it."
| > But is not OK to say it is not a bug.  It does not make GCC team look
| > serious.  Professionaly speaking. 
| 
| You can't say it is a bug across the board, it depends on the language
| definition I would say, and since long long double is an extension, there
| is no language definition to appeal to :-)

This issue it not about "long long double".

-- Gaby