«

gcc 3.4 (mingw) says error: invalid operands to binary + when I try to
to pointer arithmetic to calculate offset of a certain element.
It compiles without errors or warnings in MSVC.

Any clue why?

code (fully compilable snippet):

/* Disable MSVC W C4996 */
#ifdef _MSC_VER
#define _CRT_SECURE_NO_DEPRECATE
#endif

#include <stdio.h>
#ifndef _MSC_VER
#include <stdint.h>
#else /* _MSC_VER */
typedef unsigned __int32 uint32_t;
#endif /* _MSC_VER */

/*
* Sub-cheat (#)
*/
typedef struct sub_s {
/* Sub-cheat title */
char *label;
/* An array of 32-bit numbers containing an address or a value */
uint32_t *data;
/* The number of addresses and values */
size_t n_data;
} sub_t;

/*
* Main cheat (@)
*/
typedef struct cheat_s {
/* Cheat title */
char *label;
/* Array of sub-cheat(s) */
sub_t *sub;
/* The number of sub-cheat(s) */
size_t n_subs;
} cheat_t;

/* Global pointer to (main) cheats */
cheat_t *cheat;
/* The number of (main) cheats */
size_t n_cheats;

void foo() {
printf("debug: cheat[n_cheats - 1] @ %p\n", cheat + (cheat_t *)
(n_cheats - 1));
}

"andreyvul" <> wrote in message
news:c1751112-7eb1-4b81-a3e2-...
> gcc 3.4 (mingw) says error: invalid operands to binary + when I try to
> to pointer arithmetic to calculate offset of a certain element.
> It compiles without errors or warnings in MSVC.
>
> Any clue why?
>

<snip>

> typedef struct cheat_s {
> /* Cheat title */
> char *label;
> /* Array of sub-cheat(s) */
> sub_t *sub;
> /* The number of sub-cheat(s) */
> size_t n_subs;
> } cheat_t;
>
> /* Global pointer to (main) cheats */
> cheat_t *cheat;
> /* The number of (main) cheats */
> size_t n_cheats;
>
> void foo() {
> printf("debug: cheat[n_cheats - 1] @ %p\n", cheat + (cheat_t *)
> (n_cheats - 1));
> }
>

note that it is not valid to add two pointers together like this.
you can add a pointer and an integer, but not two pointers.

so, altering:
> printf("debug: cheat[n_cheats - 1] @ %p\n", cheat + (n_cheats -
> 1));

andreyvul <> writes:

> gcc 3.4 (mingw) says error: invalid operands to binary + when I try to
> to pointer arithmetic to calculate offset of a certain element.

[...]

> cheat_t *cheat;
[...]
> printf("debug: cheat[n_cheats - 1] @ %p\n", cheat + (cheat_t *)
> (n_cheats - 1));

You can subtract two pointers, or add an integer to a pointer,
but adding two pointers makes no sense. What are you really
trying to do?
--
char a[]="\n .CJacehknorstu";int putchar(int);int main(void){unsigned long b[]
={0x67dffdff,0x9aa9aa6a,0xa77ffda9,0x7da6aa6a,0xa6 7f6aaa,0xaa9aa9f6,0x11f6},*p
=b,i=24;for(;p+=!*p;*p/=4)switch(0[p]&3)case 0:{return 0;for(p--;i--;i--)case+
2:{i++;if(i)break;else default:continue;if(0)case 1utchar(a[i&15]);break;}}}

andreyvul wrote:
> gcc 3.4 (mingw) says error: invalid operands to binary + when I try to
> to pointer arithmetic to calculate offset of a certain element.
> It compiles without errors or warnings in MSVC.
>
> Any clue why?

Why MSVC fails to diagnose your error is something to take up with them.
The offending lines
> printf("debug: cheat[n_cheats - 1] @ %p\n", cheat + (cheat_t *)
> (n_cheats - 1));
attempts to add a pointer to a pointer, which makes no sense, even if
your cast of an int to a pointer had any portably defined meaning.
Replace it with the legal
printf("debug: cheat[n_cheats - 1] @ %p\n",
(void *) (cheat + n_cheats - 1));

Note that %p expects a pointer-to-void.

On Jan 1, 10:04 pm, Martin Ambuhl <mamb...@earthlink.net> wrote:
> andreyvul wrote:
> > gcc 3.4 (mingw) says error: invalid operands to binary + when I try to
> > to pointer arithmetic to calculate offset of a certain element.
> > It compiles without errors or warnings in MSVC.
>
> > Any clue why?
>
> Why MSVC fails to diagnose your error is something to take up with them.
> The offending lines> printf("debug: cheat[n_cheats - 1] @ %p\n", cheat + (cheat_t *)
> > (n_cheats - 1));
>
> attempts to add a pointer to a pointer, which makes no sense, even if
> your cast of an int to a pointer had any portably defined meaning.
> Replace it with the legal
> printf("debug: cheat[n_cheats - 1] @ %p\n",
> (void *) (cheat + n_cheats - 1));
>
> Note that %p expects a pointer-to-void.

My 3 A.M. C is kinda rusty. Shouldn't the integer be multiplied by
sizeof before addition?

andreyvul wrote:
> On Jan 1, 10:04 pm, Martin Ambuhl <mamb...@earthlink.net> wrote:

>> Replace it with the legal
>> printf("debug: cheat[n_cheats - 1] @ %p\n",
>> (void *) (cheat + n_cheats - 1));
>>
>> Note that %p expects a pointer-to-void.
>
> My 3 A.M. C is kinda rusty. Shouldn't the integer be multiplied by
> sizeof before addition?

No.

On Jan 1, 10:18 pm, Martin Ambuhl <mamb...@earthlink.net> wrote:
> andreyvul wrote:
> > On Jan 1, 10:04 pm, Martin Ambuhl <mamb...@earthlink.net> wrote:
> >> Replace it with the legal
> >> printf("debug: cheat[n_cheats - 1] @ %p\n",
> >> (void *) (cheat + n_cheats - 1));
>
> >> Note that %p expects a pointer-to-void.
>
> > My 3 A.M. C is kinda rusty. Shouldn't the integer be multiplied by
> > sizeof before addition?
>
> No.

So the original code was redundant AND illegal?