On Tue, May 28, 2019 at 08:27:21PM +0000, Bae, Chang Seok wrote:
> Hi all,
>
> [1] looks to have been open in quite a long period time; the size specifier has been broken obviously.
> And I can see some benefits, [2] for instance. But fixing for those a few hundreds instances has been incremental (so far).
>
> At this point, I just wonder what’s our plan, at least for this open bug. Is it open to (radically) drop
> the size specifier support?, or we need to keep fixing for those.
>
> [1] "error: mismatch in operand sizes" on scalar sse instruction, https://bugzilla.nasm.us/show_bug.cgi?id=978756
> [2] 3.7 STRICT: Inhibiting Optimization, https://www.nasm.us/xdoc/2.14.02/html/nasmdoc3.html#section-3.7
As to me -- we might need to walk over all instructions and add size
specifiers where they are missing, but this require a lot of time.
Maybe Peter has some idea?
Cyrill
|