Re: Efficient detection of signed overflow?

["Segher Boessenkool" <segher at kernel dot crashing dot org>]

>>> The comment is wrong.  The code checks for signed overflow, but the
>>> following assignment still overflwos when ux is larger than INT_MAX.
>>
>> No, it doesn't.  This conversion is implementation-defined (6.3.1.3/3),
>> and GCC does the obvious two's complement thing.  This code is fine.
>
> It's fine with GCC 4.4, and likely with GCC 4.5 as well.  But what
> about GCC 4.6?  And how will a user compiling third-party software
> notice the discrepancy (if it ever arises)?

Implementation-defined means the implementation defines
the behaviour, and GCC defines it like this:

   * `The result of, or the signal raised by, converting an integer to a
     signed integer type when the value cannot be represented in an
     object of that type (C90 6.2.1.2, C99 6.3.1.3).'

     For conversion to a type of width N, the value is reduced modulo
     2^N to be within range of the type; no signal is raised.

If this is ever to change, I'm sure you will hear about it.
Paranoid users can check the manual at every compiler release.


Segher