This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Improve PR30911 and PR31023, Ada and VRP
> Now I come to think about it, doesn't that pragma initialize everything to
> *invalid* values when possible (the impossible case is when every value is
> valid)? If so, doesn't that exactly mean you can't assume variables are in
> range when this pragma is on?
>
> PS: pragma Normalize_Scalars
I wasn't talking about that one, but about a restrictions Pragma that
exists to enforce a coding standard requiring initializations.
Normalize_Scalars is an interesting case, but one that I haven't thought
about much. But, thinking aloud, in that case we know that either the
value is in the range of the subtype or ONE SPECIFIC invalid value. That
means the issue we were talking about where the value might be different
in two places might not be able to occur, but I'm not sure.