casting args to functions...

[Doug Gwyn]

In article <115 at tdi2.UUCP> brandon at tdi2.UUCP (Brandon Allbery) writes:
>The problem is not this:  it is that ioctl() is documented as taking a pointer
>in some places and as taking an int in others.  Please note that it is not
>documented as taking a union, or is the "int" changed to a "long" on machines
>with 2-byte integers and 4-byte pointers.  Berkeley int==long-ism strikes
>again!

I don't think Berkeley int==long-ism has anything to do with this
specific problem.  If you want to roast Berkeley, how about roasting
them for deciding that those (int)s should be (int *)s and that the
size of the data structures should be encoded into an itty-bitty
field within the ioctl code.