Debugging of gcc 3.1 programs with -mfpmath=sse

[Jan Hubicka]

Hi,
while hunting down the bugs in gcc3.1 SSE scalar code generation I've noticed
that new gdb5.1 supports SSE (great :), but insist on printing the variables in
hexadecimal form, instead of as floats, that is somewhat anoying.  Since I am
having minimal insight to gdb internals, and somehow I want to keep that for
some time, I would like to ask, what needs to be done to get floats displayed
right and whether it is possible to think about it in gdb5.1.x line, as gcc 3.1
will be (as we hope) released relativly soon and it would be nice to have
SSE code generation working well.

Honza