float.h for m68k

Andreas Schwab schwab@issan.informatik.uni-dortmund.de
Tue Jun 2 01:45:00 GMT 1998


Jeffrey A Law <law@hurl.cygnus.com> writes:

|>   In message <vyzemxd4gmx.fsf@issan.informatik.uni-dortmund.de>you write:
|>   > Here is a float.h that is correct for m68k (float-i64.h contains broken
|>   > values for long double limits).  I have added float_format=m68k to all
|>   > m68k targets that do not redefine LONG_DOUBLE_TYPE_SIZE to 64.
|>   > 
|>   > 
|>   > 1998-05-28  Andreas Schwab  <schwab@issan.informatik.uni-dortmund.de>
|>   > 
|>   > 	* configure.in: Set float_format to m68k for all m68k targets that
|>   > 	do not override LONG_DOUBLE_TYPE_SIZE.
|>   > 	* config/float-m68k.h: New file.
|> Shouldn't file file be called float-i96.h since it's really the
|> 96bit IEEE format that is being used, right?

No.  The Motorola 96 bit format is different at least from the 96(80) bit
format used by Intel, not only on bit level, but LDOUBLE_MIN is different.

-- 
Andreas Schwab                                      "And now for something
schwab@issan.informatik.uni-dortmund.de              completely different"
schwab@gnu.org



More information about the Gcc-patches mailing list