>Part of the problem is that "char" may be signed on one platform, but
>unsigned on another. Wouldn't using "-fsigned-char" to GCC help? This, at
>least, would ensure the build environment is always consistent...
Yes, but this would unduly penalise those platforms where an unsigned char
is more efficient. What would actually help would be for people writing code
on i386 or whatever to compile with `-funsigned-char' once in a while and look
for extra warnings, or as Russell suggested for the compiler to be augmented
so that it warned about these constructs anyway.
It's not really any different to taking care over big/little-endian, or
32/64-bit wordsize. You just need to be aware of the dangers.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to firstname.lastname@example.org
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Apr 30 2000 - 21:00:07 EST