[Gambas-user] Possible error when doing math on bytes

[Benoit Minisini]

Le 03/01/2023 à 23:08, Brian G a écrit :
> When I do math on bytes I get an overflow now when  the value of the 
> byte goes negative
> example:
>   dim a as byte = 0
>   dim b as byte = 24
>   dim result as byte = a - b
> 
> This causes an overflow error, where in the past it was just a negative 
> number.
> Is this a bug or a new undocumented feature?
> 

Overflows were ignored before, as it is usually the case in C programs, 
even if the behaviour of overflows is unspecified. It's for speed 
reasons, or to make hacking easier, I dunno.

So I decided to detect them, for a small performance penalty.

The overflow checks are only disabled on 'Fast Unsafe' functions.

Regards,

-- 
Benoît Minisini.