[MLton] bug in new runtime: signalIsPending misalignment
Mon, 26 Jun 2006 01:31:22 -0700
> The problem is that signalIsPending is one byte and is aligned 3 mod
> 4. However, MLton thinks that signalIsPending is a word (see
> backend/rep-type.fun), and accesses it as such.
For now, I've fixed the problem by changing the declaration of
signalIsPending in the runtime to
volatile Word32_t signalIsPending;
The only remaining regression failure on Solaris is real.sml, which
exhibits a bunch of low-order-bit floating-point errors.