Re: PATCH:[darwin] fix load of a misaligned double word

[Andrew Haley <aph at redhat dot com>]

Bradley Lucier writes:
 > 
 > On Dec 31, 2003, at 1:08 PM, Andrew Haley wrote:
 > 
 > > Bradley Lucier writes:
 > >>
 > >> So it comes down to a quality of implementation question.  You're
 > >> saying that my "implementation defined" code is clearly wrong, and
 > >> everyone else's "implementation defined" conversion between pointers
 > >> and ints is OK; I have yet to be convinced.
 > >
 > > I don't understand your point.  If it is true, as I maintain, that
 > > performing arithmetic on an int* and then using the resulting pointer
 > > as a double* is undefined, then the code you present is undefined.
 > 
 > Part 7 says
 > 
 > > 7 A pointer to an object or incomplete type may be converted to a 
 > > pointer to a different object or incomplete type. If the resulting 
 > > pointer is not correctly aligned 57) for the pointed-to type, the 
 > > behavior is undefined. Otherwise, when converted back again, the 
 > > result shall compare equal to the original pointer. When a pointer to 
 > > an object is converted to a pointer to a character type, the result 
 > > points to the lowest addressed byte of the object. Successive 
 > > increments of the result, up to the size of the object, yield pointers 
 > > to the remaining bytes of the object.
 > 
 > so looking at my code.
 > 
 > int -> int * is valid, though implementation defined
 > 
 > addition of 1 to int * is valid

 > int * -> int is valid, though implementation defined
 > 
 > int -> double * is valid, though implementation defined

My reading of the standard is that the conversion from int -> double *
is only valid if that int was the result of a conversion from a
pointer that was a double*.

 > Even if I went directly from int * -> double *, it would be undefined 
 > only if it is not correctly aligned, and I know that it *is* correctly 
 > aligned.

 > > That isn't a quality of implementation issue: the code in question
 > > simply doesn't mean anything.
 > 
 > Parts 5 and 6 say that integers may be converted to and from pointers, 
 > and the results are implementation defined.  Full stop.

What it does not say, however, is that you may convert a pointer from
a double* to an int*, do arithmetic on that int* pointer, and then
convert it back to a double*.  It merely says that a double* pointer
may be converted to some other pointer and then back to a double*
pointer.

Andrew.