[PATCH] Fix up long double fphex

[Marek Polacek]

On Tue, Mar 06, 2012 at 10:26:45PM +0100, Jakub Jelinek wrote:
> I'm pretty sure this test fails on targets where long double isn't ldbl-96, but
> ldbl-128, ldbl-128ibm or dbl-64, because for those
> a will be L"0x1.88p+4" and A L"0X1.88P+4".  Only ldbl-96 %La/%LA printing ever
> prints before hexadecimal dot numbers from 2 to f/F.

Ah, I see, sorry.
Would something like this be workable?

#if LDBL_MANT_DIG == 64
  expect L"0xc.4p+1"
#else
  expect L"0x1.88p+4"
#endif
?

I've checked LDBL_MANT_DIG on PPC64 and there it is 106 (-mlong-double-128)
or 53 (-mlong-double-64).

And I guess we should run the test only when !NO_LONG_DOUBLE...

Thanks a lot for catching this.

	Marek