On Tue, 2 Oct 2001, Simon Cozens wrote:
> On Tue, Oct 02, 2001 at 10:23:22AM +0100, Simon Cozens wrote:
> Yep, all the floating point tests are failing:
>
> (gdb) print interpreter->num_reg->registers[1]
> $4 = Can't deal with a floating point number of 16 bytes.
>
> They all come out as 0.00. Dunno quite what to do about that.
At least some places previously assumed that sizeof(nv) == 2*sizeof(iv). I
changed at least some of those to use the ratio nvsize/ivsize from
PConfig, but that's not quite right either, since the ratio isn't
guaranteed to be an integer. A long double may well be 12 bytes long and
an iv may be 8 bytes long.
I've just re-synced and can poke around a bit today, but that's one place
I'd start looking.
--
Andy Dougherty [EMAIL PROTECTED]
Dept. of Physics
Lafayette College, Easton PA 18042
