pointer-in-array test

[Dik T. Winter]

Re: pointer-in-array test

In article <1133145.9pkQ6i reiI@holzmayer. ifr.rt> holzmayer.bernh ard@deadspam.co m writes:
....[color=blue]
> What about this solution:
> int ptrInArrSafe(T *p,T *a,int max)
> /* check whether p points to an element of a[max] */
> {
> long px = (long)p;
> long ax = (long)a;
> if (px<ax) return 0; /*out*/
> if (px> (long)&a[max-1]) return 0; /* out */
> if ( (px-ax) % sizeof(T)) return 0; /* not at element start*/
> return 1; /* in and on element start */
> }
>
> compiles, links, runs without problems with my gcc.
> And, as far as I remember, it's legal, to compare pointers after
> conversion to long.[/color]

But the result is implementation defined. I know of at least one
system where it will fail to give the correct answer, then T is char.
For instance:
(long)(a + 7) > (long)(a + 8)
when a is an array of type char.
--
dik t. winter, cwi, kruislaan 413, 1098 sj amsterdam, nederland, +31205924131
home: bovenover 215, 1025 jn amsterdam, nederland; http://www.cwi.nl/~dik/

[Eric Sosman]

Re: pointer-in-array test

Ike Naar wrote:[color=blue]
> [...]
> A more efficient algorithm is
>
> int ptrInArrUnsafe( T *p,T *a,int max)
> /* check whether p points to an element of a[max] */
> {
> int j=p-a; /* possibly undefined */
> return 0<=j && j<max && a+j==p;
> }
>
> If p points to an element of a[max], j will be a valid index and
> the return expression will evaluate to true.[/color]

Agree.
[color=blue]
> If p does not point to an element of a[max], the pointer subtraction may
> invoke undefined behaviour; garbage will be stored in j, and at least one
> of the conjuncts of the return expression will evaluate to false.[/color]

Disagree. Everything seems correct up to the semicolon,
but after that things fall apart:

- First, it is not necessarily the case that garbage will
be stored in j. "Undefined behavior" is "undefined, "
and does not necessarily result in storing anything at
all anywhere at all.

- Second, even if something does get stored into j, it's
not certain that the "something" will be recognizable
as "garbage." For example, the value zero might be
stored into j, just as if you had called

ptrInArrUnsafe( &array[0], array, 1)

- ... and if this (or something like it) happens, the
return expression will evaluate true, not false.
[color=blue]
> Apart from storing garbage in j, the pointer subtraction could, in theory,
> crash the system, re-format the floppy disk or wake up nasal daemons,
> but this has not happened so far on any platform with any compiler
> (tried several).[/color]

Personally, I've never seen demons fly from my nose (they're
awfully fast, and are usually out of sight before I notice them).
But one need not invoke exotic architectures to get weird behavior
out of violating the rules of pointer subtraction. For example,
consider a "plain vanilla" system with nice linear addressing,
where pointers are just unadorned numeric addresses. Now try
this on for size:

typedef char[8] T;
T array[5];
printf ("%d\n", ptrInArrUnsafe( &array[0][1], array, 5));

It is quite likely (not certain, of course: undefined behavior
is undefined) that j will compute as zero inside the function,
leading the function to conclude that the pointer is valid.
The conclusion is utterly wrong, of course: the pointer aims
not at the start of a T object, but one byte past the start
of the first T object in the array.
[color=blue]
> My question is, do you experts in this group see any real problem with
> the second algorithm?[/color]

Yes, as explained above.
[color=blue]
> Alternatively, would there be a faster algorithm
> than the first one, that does not invoke undefined behaviour?[/color]

I cannot think of one. As CBF suggests, you should (a)
re-examine whether such a test is required, and if it is
you should (b) isolate the test to a system-dependent source
file, not intended to be portable.

--
Eric.Sosman@sun .com

[CBFalconer]

Re: pointer-in-array test

Ike Naar wrote:[color=blue]
>[/color]
.... snip ...[color=blue]
>
> But apart from stylistic differences, I see no reason why your
> solution would be any faster or slower than mine.
> Both algorithms perform a linear search on the array.[/color]

I did say "possibly". It would depend on the optimization
capabilities of the compiler.

--
Chuck F (cbfalconer@yah oo.com) (cbfalconer@wor ldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home .att.net> USE worldnet address!

[Ike Naar]

Re: pointer-in-array test

Hi Eric,

Eric Sosman <Eric.Sosman@su n.com> wrote:
: Ike Naar wrote:
:> int ptrInArrUnsafe( T *p,T *a,int max)
:> /* check whether p points to an element of a[max] */
:> {
:> int j=p-a; /* possibly undefined */
:> return 0<=j && j<max && a+j==p;
:> }
:> [...]
:> If p does not point to an element of a[max], the pointer subtraction may
:> invoke undefined behaviour; garbage will be stored in j, and at least one
:> of the conjuncts of the return expression will evaluate to false.
: Disagree. Everything seems correct up to the semicolon,
: but after that things fall apart:
: - First, it is not necessarily the case that garbage will
: be stored in j. "Undefined behavior" is "undefined, "
: and does not necessarily result in storing anything at
: all anywhere at all.

Of course the implementation is free to do anything it wants after
undefined behaviour, but let's pretend that the damage is limited
to whatever is stored in j. If nothing is stored in j, fine. Perhaps
j retains its old value (whatever that was).

: - Second, even if something does get stored into j, it's
: not certain that the "something" will be recognizable
: as "garbage."

It seems that our notions of "garbage" differ.
My garbage can have any value.

: For example, the value zero might be
: stored into j, just as if you had called
: ptrInArrUnsafe( &array[0], array, 1)
: - ... and if this (or something like it) happens, the
: return expression will evaluate true, not false.

If 0<=j<max, the last conjuct (a+j==p) has to be false, due to the
initial assumption that p was not pointing to an element of a .

Conversely, if the return expression evaluates true there would
have been no undefined behaviour, since in that case there exists
an index j such that 0<=j<max and &a[j]==p (that's what the return
expression implies), i.e. p points to an element of a, and
the pointer subtraction (p-a) would have been OK.

:> [...]
: But one need not invoke exotic architectures to get weird behavior
: out of violating the rules of pointer subtraction. For example,
: consider a "plain vanilla" system with nice linear addressing,
: where pointers are just unadorned numeric addresses. Now try
: this on for size:

: typedef char[8] T;

<nitpicking -- this is C.L.C. after all, isn't it?>
typedef char T[8];
</nitpicking>

: T array[5];
: printf ("%d\n", ptrInArrUnsafe( &array[0][1], array, 5));

This is an extra complication that does not apply to the original problem
where p is supposed to have type T* . &array[0][1] is not a T* .

: It is quite likely (not certain, of course: undefined behavior
: is undefined) that j will compute as zero inside the function,
: leading the function to conclude that the pointer is valid.
: The conclusion is utterly wrong, of course: the pointer aims
: not at the start of a T object, but one byte past the start
: of the first T object in the array.

: --
: Eric.Sosman@sun .com

--
mail to ike at iae dot nl

[Eric Sosman]

Re: pointer-in-array test

Ike Naar wrote:[color=blue]
>
> Hi Eric,
>
> Eric Sosman <Eric.Sosman@su n.com> wrote:
> : Ike Naar wrote:
> :> int ptrInArrUnsafe( T *p,T *a,int max)
> :> /* check whether p points to an element of a[max] */
> :> {
> :> int j=p-a; /* possibly undefined */
> :> return 0<=j && j<max && a+j==p;
> :> }
> :>
> : - ... and if this (or something like it) happens, the
> : return expression will evaluate true, not false.
>
> If 0<=j<max, the last conjuct (a+j==p) has to be false, due to the
> initial assumption that p was not pointing to an element of a .[/color]

Ah, yes. Sorry about that; my reading comprehension
diminishes rapidly once the U.B. alarm bells start ringing.
[color=blue]
> Conversely, if the return expression evaluates true there would
> have been no undefined behaviour, since in that case there exists
> an index j such that 0<=j<max and &a[j]==p (that's what the return
> expression implies), i.e. p points to an element of a, and
> the pointer subtraction (p-a) would have been OK.
>
> :> [...]
> : But one need not invoke exotic architectures to get weird behavior
> : out of violating the rules of pointer subtraction. For example,
> : consider a "plain vanilla" system with nice linear addressing,
> : where pointers are just unadorned numeric addresses. Now try
> : this on for size:
>
> : typedef char[8] T;
>
> <nitpicking -- this is C.L.C. after all, isn't it?>
> typedef char T[8];
> </nitpicking>[/color]

Right again, and I've clearly become Java-polluted.
[color=blue]
> : T array[5];
> : printf ("%d\n", ptrInArrUnsafe( &array[0][1], array, 5));
>
> This is an extra complication that does not apply to the original problem
> where p is supposed to have type T* . &array[0][1] is not a T* .[/color]

Right again, but this time it was just an editing error.
I'd started with

(T*)((char*)arr ay + 1)

.... and lost the cast when simplifying. The first argument
ought to have been

(T*)&array[0][1]

Now, there are lots of opportunities for things to go wrong
with this. The conversion to T* might garble the address --
but in the example I was assuming a "plain vanilla" machine
on which that wouldn't happen. The pointer subtraction in
the function is still likely to yield zero by (more or less)
subtracting the two addresses to get 1 and then right-shifting
by three bits. It is, I guess, unlikely that a "plain vanilla"
machine would decide that (T*)&array[0][1] and array[0]+0 are
equal -- no guarantees, of course, but unlikely.

All in all, I'd say my attempt to express objections has
turned into an awful hash. Nonetheless ("Don't confuse me
with the facts; my mind is made up!") I'll still second CBF's
recommendation that shenanigans of this sort should be avoided,
or at the very least isolated.

--
Eric.Sosman@sun .com

[Christian Bau]

Re: pointer-in-array test

In article <1133145.9pkQ6i reiI@holzmayer. ifr.rt>,
Bernhard Holzmayer <holzmayer.bern hard@deadspam.c om> wrote:
[color=blue]
> Sorry,
> pardon me for not checking my code intensively enough.
> I just checked with a 'gcc -pedantic' which didn't throw an error :(
>
> What about this solution:
> int ptrInArrSafe(T *p,T *a,int max)
> /* check whether p points to an element of a[max] */
> {
> long px = (long)p;
> long ax = (long)a;
> if (px<ax) return 0; /*out*/
> if (px> (long)&a[max-1]) return 0; /* out */
> if ( (px-ax) % sizeof(T)) return 0; /* not at element start*/
> return 1; /* in and on element start */
> }
>
> compiles, links, runs without problems with my gcc.
> And, as far as I remember, it's legal, to compare pointers after
> conversion to long.[/color]

It just isn't guaranteed to produce any meaningful result.

[Keith Thompson]

Re: pointer-in-array test

Bernhard Holzmayer <holzmayer.bern hard@deadspam.c om> writes:[color=blue]
> Keith Thompson wrote:
>[color=green]
> > Barry Schwarz <schwarzb@deloz .net> writes:
> > [...][color=darkred]
> >> If your system supports the optional intptr_t type, you can
> >> convert p, a, and a+1 to integers and determine the correct
> >> answer without fear of undefined behavior.[/color]
> >
> > The only guarantee for intptr_t is that you can convert a void* to
> > intptr_t and back to void*, and the result will compare equal to
> > the
> > original void*. It's not clear (to me) that relational operators
> > are sufficiently well-defined to allow reliably checking whether a
> > pointer points within an array.
> >
> > Suppose a char* pointer consists of a word pointer with a byte
> > offset in the high-order bits, and that pointer-to-integer[/color]
> pointer-to-integer is not enough.
> IMHO you have to use pointer-to-long which must be well-defined
> according to the standard, which means, that a long is big enough
> to hold a pointer on that system. An integer might be too short.[/color]

By "word pointer", I meant a pointer to a word. There's no necessary
relationship between a integer and a pointer-to-integer -- and the
term "integer" includes all integer types, not just int.

The concrete example I have in mind is Cray vector systems. A machine
word is 64 bits; there are no instructions that operate directly on
8-bit bytes. Nevertheless, the C implementation has CHAR_BIT==8, so
it needs to have a way to represent pointers to 8-bit bytes.

An int* (int is 64 bits) is simply a word pointer with the obvious
representation. If you take an int*, treat it as a 64-bit integer,
increment it, and convert it back to int*, the resulting pointer will
point to the 64-bit word immediately following the original one.

A char* consists of a word pointer, pointing to the 64-bit word
containing the byte, with a 3-bit byte offset (value 0..7) stored in
the high-order 3 bits (which would otherwise be zero, since no such
system has a large enough address space to need them).

--
Keith Thompson (The_Other_Keit h) kst-u@mib.org <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
Schroedinger does Shakespeare: "To be *and* not to be"

[Bernhard Holzmayer]

Re: pointer-in-array test

Dik T. Winter wrote:
[color=blue]
> In article <1133145.9pkQ6i reiI@holzmayer. ifr.rt>
> holzmayer.bernh ard@deadspam.co m writes: ...[color=green]
> > What about this solution:
> > int ptrInArrSafe(T *p,T *a,int max)
> > /* check whether p points to an element of a[max] */
> > {
> > long px = (long)p;
> > long ax = (long)a;
> > if (px<ax) return 0; /*out*/
> > if (px> (long)&a[max-1]) return 0; /* out */
> > if ( (px-ax) % sizeof(T)) return 0; /* not at element
> > start*/
> > return 1; /* in and on element start */
> > }
> >
> > compiles, links, runs without problems with my gcc.
> > And, as far as I remember, it's legal, to compare pointers
> > after conversion to long.[/color]
>
> But the result is implementation defined. I know of at least one
> system where it will fail to give the correct answer, then T is
> char. For instance:
> (long)(a + 7) > (long)(a + 8)
> when a is an array of type char.[/color]

That's not the same.
Would you please try the following:
(long)(&a[7]) < (long)(&a[8])

this should give the correct result.

Reason for your strange behaviour is possibly the implicite type
cast, which converts a before adding.
Instead, the later version increments a as a pointer and then
converts it.

Bernhard

[Bernhard Holzmayer]

Re: pointer-in-array test

Christian Bau wrote:
[color=blue]
> In article <1133145.9pkQ6i reiI@holzmayer. ifr.rt>,
> Bernhard Holzmayer <holzmayer.bern hard@deadspam.c om> wrote:
>[color=green]
>> Sorry,
>> pardon me for not checking my code intensively enough.
>> I just checked with a 'gcc -pedantic' which didn't throw an error
>> :(
>>
>> What about this solution:
>> int ptrInArrSafe(T *p,T *a,int max)
>> /* check whether p points to an element of a[max] */
>> {
>> long px = (long)p;
>> long ax = (long)a;
>> if (px<ax) return 0; /*out*/
>> if (px> (long)&a[max-1]) return 0; /* out */
>> if ( (px-ax) % sizeof(T)) return 0; /* not at element
>> start*/
>> return 1; /* in and on element start */
>> }
>>
>> compiles, links, runs without problems with my gcc.
>> And, as far as I remember, it's legal, to compare pointers after
>> conversion to long.[/color]
>
> It just isn't guaranteed to produce any meaningful result.[/color]

Why. Please give arguments.

Bernhard

[Christian Bau]

Re: pointer-in-array test

In article <lnptajglui.fsf @nuthaus.mib.or g>,
Keith Thompson <kst-u@mib.org> wrote:
[color=blue]
> By "word pointer", I meant a pointer to a word. There's no necessary
> relationship between a integer and a pointer-to-integer -- and the
> term "integer" includes all integer types, not just int.
>
> The concrete example I have in mind is Cray vector systems. A machine
> word is 64 bits; there are no instructions that operate directly on
> 8-bit bytes. Nevertheless, the C implementation has CHAR_BIT==8, so
> it needs to have a way to represent pointers to 8-bit bytes.
>
> An int* (int is 64 bits) is simply a word pointer with the obvious
> representation. If you take an int*, treat it as a 64-bit integer,
> increment it, and convert it back to int*, the resulting pointer will
> point to the 64-bit word immediately following the original one.
>
> A char* consists of a word pointer, pointing to the 64-bit word
> containing the byte, with a 3-bit byte offset (value 0..7) stored in
> the high-order 3 bits (which would otherwise be zero, since no such
> system has a large enough address space to need them).[/color]

Take a 16 bit x86 system, using the "huge" memory model: A pointer
consists of 16 bit segment and 16 bit offset. When the offset exceeds
65535, it wraps around to 0 and the segment is increased by eight. So it
can happen that

(unsigned long) p == 0x3241ffff
(unsigned long) (p+1) == 0x32490000

Doing pointer arithmetic on unsigned long would prove fatally wrong. In
the case above, if p and p+1 are both valid pointers to char objects,
then

* (char *) (((unsigned long) p) + 1) = '\0';

might crash your program. It will definitely not write to p [1].

[Bernhard Holzmayer]

Re: pointer-in-array test

Thomas Stegen wrote:
[color=blue]
> Bernhard Holzmayer wrote:
>[color=green]
>> Sorry,
>> pardon me for not checking my code intensively enough.
>> I just checked with a 'gcc -pedantic' which didn't throw an error
>> :(
>>
>> What about this solution:
>> int ptrInArrSafe(T *p,T *a,int max)
>> /* check whether p points to an element of a[max] */
>> {
>> long px = (long)p;
>> long ax = (long)a;
>> if (px<ax) return 0; /*out*/
>> if (px> (long)&a[max-1]) return 0; /* out */
>> if ( (px-ax) % sizeof(T)) return 0; /* not at element
>> start*/
>> return 1; /* in and on element start */
>> }[/color][/color]

After a close look to K&R 5.3 and additional sources, I'd like
to improve the above a little. What about the following modified
version:

int ptrInArrSafe(T *p,T *a,int max)
/* check whether p points to an element of a[max] */
{
long index;

/* as long as a points to a valid array, this is a legal
test for the lower boundary*/
if (p<a) return 0; /*out */
/* evaluation of element's address is valid for all elements
including the virtual last+1 element, though its content may not be
retrieved, comparison is legal test for upper boundary */
if (!(p<&a[max])) return 0; /* out */

/* now we can be sure that we are inside array, because
standard requires that array is in a logically "dense" order, from
point of logical addresses they must be accessible in ascending
order */

index = (long)(p - a);
/* as far as I understand, it's possible that p-a returns
the difference in multiples of sizeof(T), so that if element size
were 2, a[2]-a[1] would give 1 instead of 2.
Although I never found a compiler doing this, I replace sizeof(T) by
a method which should work in both cases. */
if ( (p-a) % (&a[1]-&a[0]))
return 0; /* not at element boundary*/

return 1; /* in and on element boundary */
}

Bernhard

[Irrwahn Grausewitz]

Re: pointer-in-array test

Bernhard Holzmayer <holzmayer.bern hard@deadspam.c om> wrote:[color=blue]
>Christian Bau wrote:[color=green]
>> Bernhard Holzmayer <holzmayer.bern hard@deadspam.c om> wrote:[/color][/color]
<snip>[color=blue][color=green][color=darkred]
>>> What about this solution:
>>> int ptrInArrSafe(T *p,T *a,int max)
>>> /* check whether p points to an element of a[max] */
>>> {
>>> long px = (long)p;
>>> long ax = (long)a;[/color][/color][/color]
<snip>[color=blue][color=green][color=darkred]
>>> compiles, links, runs without problems with my gcc.
>>> And, as far as I remember, it's legal, to compare pointers after
>>> conversion to long.[/color]
>>
>> It just isn't guaranteed to produce any meaningful result.[/color]
>
>Why. Please give arguments.[/color]

As has already been pointed out upthread (e.g. by pete in message
<4073D355.5C39@ mindspring.com> ), it's implementation-defined if
pointer values are meaningfully representable in an object of _any_
integral type at all.

HTH
Regards
--
Irrwahn Grausewitz (irrwahn33@free net.de)
welcome to clc: http://www.ungerhu.com/jxh/clc.welcome.txt
clc faq-list : http://www.faqs.org/faqs/C-faq/faq/
clc OT guide : http://benpfaff.org/writings/clc/off-topic.html

[Bernhard Holzmayer]

Re: pointer-in-array test

pete wrote:
[color=blue]
> Bernhard Holzmayer wrote:
>[color=green]
>> IMHO you have to use pointer-to-long which must be well-defined
>> according to the standard, which means, that a long is big enough
>> to hold a pointer on that system. An integer might be too short.[/color]
>
> Whether or not a pointer can be converted to any integer type
> is implementation defined.
> The standard does not guarantee that relationships which hold
> for two pointers, will also hold for their converted integer
> values.
>
> N869
> 6.3.2.3 Pointers
> [#6] Any pointer type may be converted to an integer
> [#type.
> Except as previously specified, the result
> is
> implementation-defined. If the result cannot be
> represented
> in the integer type, the behavior is undefined. The
> result need not be in the range of values of any integer
> type.
>[/color]

1) Isn't it correct, that long has to be of the size to hold any
pointer, so that a conversion:
(pType *) --> (void *) ---> (long)
and vice versa is legal?

2) Isn't it correct, that arrays must have their elements arranged
subsequently, so that array == &array[0] and
&array[1]>&array[0] are valid?

[Bernhard Holzmayer]

Re: pointer-in-array test

Bernhard Holzmayer wrote:
[color=blue]
> Thomas Stegen wrote:
>[color=green]
>> Bernhard Holzmayer wrote:
>>[color=darkred]
>>> Sorry,
>>> pardon me for not checking my code intensively enough.
>>> I just checked with a 'gcc -pedantic' which didn't throw an
>>> error
>>> :(
>>>
>>> What about this solution:
>>> int ptrInArrSafe(T *p,T *a,int max)
>>> /* check whether p points to an element of a[max] */
>>> {
>>> long px = (long)p;
>>> long ax = (long)a;
>>> if (px<ax) return 0; /*out*/
>>> if (px> (long)&a[max-1]) return 0; /* out */
>>> if ( (px-ax) % sizeof(T)) return 0; /* not at element
>>> start*/
>>> return 1; /* in and on element start */
>>> }[/color][/color]
>
> After a close look to K&R 5.3 and additional sources, I'd like
> to improve the above a little. What about the following modified
> version:
>
> int ptrInArrSafe(T *p,T *a,int max)
> /* check whether p points to an element of a[max] */
> {
> long index;[/color]
ignore above line![color=blue]
>
> /* as long as a points to a valid array, this is a legal
> test for the lower boundary*/
> if (p<a) return 0; /*out */
> /* evaluation of element's address is valid for all
> elements
> including the virtual last+1 element, though its content may not
> be retrieved, comparison is legal test for upper boundary */
> if (!(p<&a[max])) return 0; /* out */
>
> /* now we can be sure that we are inside array, because
> standard requires that array is in a logically "dense" order, from
> point of logical addresses they must be accessible in ascending
> order */
>
> index = (long)(p - a);[/color]
ignore above line![color=blue]
> /* as far as I understand, it's possible that p-a returns
> the difference in multiples of sizeof(T), so that if element size
> were 2, a[2]-a[1] would give 1 instead of 2.
> Although I never found a compiler doing this, I replace sizeof(T)
> by a method which should work in both cases.[/color]
Keith, would this hold on your Cray with the strange CHAR
representation, that p-a still returns a reasonable result?
*/[color=blue]
> if ( (p-a) % (&a[1]-&a[0]))
> return 0; /* not at element boundary*/
>
> return 1; /* in and on element boundary */
> }
>
> Bernhard[/color]

[Irrwahn Grausewitz]

Re: pointer-in-array test

Bernhard Holzmayer <holzmayer.bern hard@deadspam.c om> wrote:
<snip>[color=blue]
>After a close look to K&R 5.3 and additional sources, I'd like
>to improve the above a little. What about the following modified
>version:
>
> int ptrInArrSafe(T *p,T *a,int max)
> /* check whether p points to an element of a[max] */
> {
> long index;
>
> /* as long as a points to a valid array, this is a legal
> test for the lower boundary*/
> if (p<a) return 0; /*out */[/color]

If p and a didn't point into (or one past) the same array in
the first place, you already invoked undefined behaviour here
(C99 6.5.8#5). All bets off.
[color=blue]
> /* evaluation of element's address is valid for all elements
>including the virtual last+1 element, though its content may not be
>retrieved, comparison is legal test for upper boundary */
> if (!(p<&a[max])) return 0; /* out */[/color]

Ditto.
[color=blue]
> /* now we can be sure that we are inside array, because
>standard requires that array is in a logically "dense" order, from
>point of logical addresses they must be accessible in ascending
>order */
>
> index = (long)(p - a);[/color]

Forget about the cast. Declare index as ptrdiff_t instead.
[color=blue]
> /* as far as I understand, it's possible that p-a returns
>the difference in multiples of sizeof(T), so that if element size
>were 2, a[2]-a[1] would give 1 instead of 2.[/color]

That's not only possible, it's /required/ for a conforming
implementation to behave like this (C99 6.5.6#9): pointer
subtraction always yields values in units of element size.
[color=blue]
>Although I never found a compiler doing this,[/color]

Then you only found non-conforming compilers up until now.
[color=blue]
>I replace sizeof(T) by
>a method which should work in both cases. */
> if ( (p-a) % (&a[1]-&a[0]))[/color]

Since &a[1]-&a[0] (==a+1-a) always yields 1, the condition
is always false.
[color=blue]
> return 0; /* not at element boundary*/
>
> return 1; /* in and on element boundary */
> }[/color]

End of the story: if you ever need to verify if a specific
pointed-to object is part of a certain array, carry around
the array length plus indices and compare these instead of
some 'raw' pointers. And, IMHO, if one feels the need to
perform the operation desired by the OP, one should first
rethink the algorithm, because it's most probably broken by
design.

HTH
Regards
--
Irrwahn Grausewitz (irrwahn33@free net.de)
welcome to clc: http://www.ungerhu.com/jxh/clc.welcome.txt
clc faq-list : http://www.faqs.org/faqs/C-faq/faq/
clc OT guide : http://benpfaff.org/writings/clc/off-topic.html