hipl-core team mailing list archive

[Diego Biurrun <diego@xxxxxxxxxx>]

On Thu, Sep 16, 2010 at 04:32:15PM +0200, Christof Mroz wrote:
> On Thu, 16 Sep 2010 15:25:22 +0200, René Hummen  
>
>> +    const int high_quad = (c & 0xF0) >> 4;
>> +    const int low_quad = c & 0x0F;
>
> OK but compilers figure out c / 0x10 and c % 0x10 nowadays AFAIK.

In my experience all assumptions about compilers being smart, much less
"nowadays" are wrong.  Until you show us the generated assembly I
consider your statement as false :)

Diego