Re: [IDEA+RFC] Possible solution for min()/max() war

[Petr Vandrovec]

On  4 Sep 01 at 12:09, VDA wrote:
> min2 performs a very strict checking.
> min3 gives you full control, but checks for too small target type.
> Do anybody see any flaws? Any improvements?
> For compiler folks: Why GCC compiles ...f(1);f(1);f(1)... to
> ...
>         pushl $1
>         call f
>         addl $32,%esp  <-- My grandma optimizes better
>         addl $-12,%esp <-- ?
>         pushl $1
>         call f
>         addl $16,%esp  <-- ?
>         addl $-12,%esp <-- ?

Compile with 'gcc -mpreferred-stack-boundary=2', gcc generates
unbelievable stupid code for other values.
                                        Best regards,
                                            Petr Vandrovec
                                            [e-mail]