4 ms·

ptrdiff_t is a signed type - after all, the diff stands for difference, which implies that it could be negative. size_t is an unsigned type - it is used to rep

ptrdiff_t is a signed type - after all, the diff stands for difference, which implies that it could be negative. size_t is an unsigned type - it is used to represent the sizes of objects in memory, which can never be negative.

0x0 11y ago
But couldn't you just get along with unsigned types for both (i.e. just size_t)? I mean, two's complement means adding or subtracting is the same for unsigned and signed ints? Is it just for the convenience of being able to compare along the lines of (somePtrDiff < 0)? But then again you could just as well compare (somePtr > someOtherPtr)? Another funny thing I just discovered: You can't declare an array like "char somearray[SIZE_MAX]"; clang gives "error: array is too large (18446744073709551615 elements)". But you can use SIZE_MAX as an array index (although it gives a warning "array index -1 is past the end of the array"). Also, trying to declare an array like "char somearray[SIZE_MAX/8]" inside a simple main() just sends clang into 100% cpu...
- ben0x539 11y ago
  Sometimes you actually want to do math beyond adding and ubstracting with ptrdiff_t. Multiply a potentially negative offset with something? Compare two offsets, one of them negative, with each other?
  - 0x0 11y ago
    For multiplication, I think the math works out the same for unsigned and signed, within mod bit-size and you throw away the "overflow"? I.e. (for 16bit pointers), 0xff_ff * 2 == 0x1_ff_fe (-1*2 == -2). For greater-than/smaller-than comparisons, I see the utility of working in the signed space though. I guess it all depends on expectations. If you have a pointer to the very last byte of a 2^32 address space (0xFFFF_FFFF), and subtract a pointer to the second byte from the start (0x0000_0001), you'll end up with a diff that should be "SIZE_MAX-2", but will be parsed as "-2" in the signed interpretation.
    - mikeash 11y ago
      Pointer subtraction is only valid for pointers within the same object. To overflow a signed ptrdiff_t, you'd need to allocate an array of char with more than 2^31 elements. This can be done on some systems, but it's really rare.
    - ben0x539 11y ago
      > For multiplication, I think the math works out the same for unsigned and signed, within mod bit-size and you throw away the "overflow"? Yeah, true. I recalled there being separate instructions and then messed up my napkin math to check if it made a difference within the relevant range.

0x0 11y ago

But couldn't you just get along with unsigned types for both (i.e. just size_t)? I mean, two's complement means adding or subtracting is the same for unsigned and signed ints? Is it just for the convenience of being able to compare along the lines of (somePtrDiff < 0)? But then again you could just as well compare (somePtr > someOtherPtr)? Another funny thing I just discovered: You can't declare an array like "char somearray[SIZE_MAX]"; clang gives "error: array is too large (18446744073709551615 elements)". But you can use SIZE_MAX as an array index (although it gives a warning "array index -1 is past the end of the array"). Also, trying to declare an array like "char somearray[SIZE_MAX/8]" inside a simple main() just sends clang into 100% cpu...

ben0x539 11y ago
Sometimes you actually want to do math beyond adding and ubstracting with ptrdiff_t. Multiply a potentially negative offset with something? Compare two offsets, one of them negative, with each other?
- 0x0 11y ago
  For multiplication, I think the math works out the same for unsigned and signed, within mod bit-size and you throw away the "overflow"? I.e. (for 16bit pointers), 0xff_ff * 2 == 0x1_ff_fe (-1*2 == -2). For greater-than/smaller-than comparisons, I see the utility of working in the signed space though. I guess it all depends on expectations. If you have a pointer to the very last byte of a 2^32 address space (0xFFFF_FFFF), and subtract a pointer to the second byte from the start (0x0000_0001), you'll end up with a diff that should be "SIZE_MAX-2", but will be parsed as "-2" in the signed interpretation.
  - mikeash 11y ago
    Pointer subtraction is only valid for pointers within the same object. To overflow a signed ptrdiff_t, you'd need to allocate an array of char with more than 2^31 elements. This can be done on some systems, but it's really rare.
  - ben0x539 11y ago
    > For multiplication, I think the math works out the same for unsigned and signed, within mod bit-size and you throw away the "overflow"? Yeah, true. I recalled there being separate instructions and then messed up my napkin math to check if it made a difference within the relevant range.