Why is targetm.fixed_condition_code_regs a function?

[Zack Weinberg]

Richard Henderson <rth@redhat.com> writes:

> On Wed, Feb 18, 2004 at 03:15:55PM -0800, Zack Weinberg wrote:
>> I don't think it is possible for this to change at runtime...
>
> I thought maybe it might change based on -march switches,
> particularly the presence or absence of the second register.

Hmm, I suppose, but couldn't we deal with that by modifying the value
in the array in override_options?  Targets already do plenty of
diddling of the targetm structure at that point.

zw