[Bug c/102924] New: incorrect output for error "cannot mix operands of decimal floating and other floating types"

[Bug c/102924] New: incorrect output for error "cannot mix operands of decimal floating and other floating types"

[vincent-gcc at vinc17 dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102924

            Bug ID: 102924
           Summary: incorrect output for error "cannot mix operands of
                    decimal floating and other floating types"
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vincent-gcc at vinc17 dot net
  Target Milestone: ---

On the following C code

void foo (int i)
{
  (void) (((float) 0 + (_Decimal64) 0) + ((float) 0 + (_Decimal64) 0));
  (void) ((i ? (float) 0 : (_Decimal64) 0) == (i ? (float) 0 : (_Decimal64)
0));
  (void) ((i ? (int *) 0 : (long *) 0) == (i ? (int *) 0 : (long *) 0));
}

GCC gives

tst.c: In function 'foo':
tst.c:3:3: error: cannot mix operands of decimal floating and other floating
types
    3 |   (void) (((float) 0 + (_Decimal64) 0) + ((float) 0 + (_Decimal64) 0));
      |   ^
tst.c:3:3: error: cannot mix operands of decimal floating and other floating
types
tst.c:4:3: error: cannot mix operands of decimal floating and other floating
types
    4 |   (void) ((i ? (float) 0 : (_Decimal64) 0) == (i ? (float) 0 :
(_Decimal64) 0));
      |   ^
tst.c:4:3: error: cannot mix operands of decimal floating and other floating
types
tst.c:5:26: warning: pointer type mismatch in conditional expression
    5 |   (void) ((i ? (int *) 0 : (long *) 0) == (i ? (int *) 0 : (long *)
0));
      |                          ^
tst.c:5:58: warning: pointer type mismatch in conditional expression
    5 |   (void) ((i ? (int *) 0 : (long *) 0) == (i ? (int *) 0 : (long *)
0));
      |                                                          ^

For the error "cannot mix operands of decimal floating and other floating
types", there are two issues (possibly related):

1. The caret location is incorrect: at the beginning of the full expression
instead of somewhere in the concerned subexpression; I would expect something
like the output for pointer type mismatch in conditional expression.

2. The detail for the second error in the expression is missing. With the caret
location issue, the detail would be the same, so that this isn't much a
problem, but once (1) is fixed, it would be different.

Tested from GCC 8 to gcc (Debian 20210918-1) 12.0.0 20210918 (experimental)
[master r12-3644-g7afcb534239].