TR1 testing mosh pit

[Benjamin Kosnik]

Hey Ed, thanks for this patch. I believe it solves the following issue 
pointed out by Paolo previously:

1) __isnan overloads need to be inlined.

In addition, I have fixed up the remaining issues that Paolo deemed 
blockers.

2) --disable-c99. This was somewhat problematic, because 
poly_laguerre.tcc does some strange things about checking __isnan on a 
template argument that is non-floating point, and because of lacking 
testsuite infrastructure to disable c99 tests when the target doesn't 
support them. Please find both attached as patches.

The poly_laguerre.tcc patch is minimal, but effective. I believe the 
better long term fix is to remove _Tpa as a template parameter, since it 
is always (as far as I can see) unsigned int. However, perhaps I'm 
missing something: you can figure out what you'd like to do after this 
goes in.

I've tested my patches overlayed with your last patch on:

x86/linux
x86_64/linux
powerpc64/linux

No fails. Great job. I feel as if this meets Paolo's last criteria, that 
  this be stable and unproblematic.

So, I'd like you to check this in, preferably before the weekend. If you 
are unable to do so, or would rather me do it, provide me with a 
ChangeLog entry that is current and I'll happily do it for you.

Again, thanks for this important work.

best,
benjamin
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: p.20070315-poly_laguerre
URL: <http://gcc.gnu.org/pipermail/libstdc++/attachments/20070315/722942ad/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: p.20070315-require-c-std
URL: <http://gcc.gnu.org/pipermail/libstdc++/attachments/20070315/722942ad/attachment-0001.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: CL
URL: <http://gcc.gnu.org/pipermail/libstdc++/attachments/20070315/722942ad/attachment-0002.ksh>