[Bug debug/63572] [5 Regression] ICF breaks user debugging experience

["jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63572

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-10-17
                 CC|                            |ccoutant at gcc dot gnu.org,
                   |                            |jan.kratochvil at redhat dot com,
                   |                            |jason at gcc dot gnu.org,
                   |                            |mark at gcc dot gnu.org
   Target Milestone|---                         |5.0
     Ever confirmed|0                           |1

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
For the different locations and blocks, I meant something like:

struct S { int a; int b; int c; };
struct T { int d; int e; int f; };

static int
f1 (struct S *x)
{
  int g = x->a * 7;
  {
    int h = x->b * 11;
    {
      int i = x->c;
      return g + h + i;
    }
  }
}

static int
f2 (struct T *x)
{
  int j = x->d * 7; int k = x->e * 11;
  int l = x->f;
  return j + k + l;
}

int f3 (struct S *x) { return f1 (x); }
int f4 (struct T *x) { return f2 (x); }
int f5 (struct S *x) { return f1 (x) + 1; }
int f6 (struct T *x) { return f2 (x) + 1; }

(but for some reason ICF doesn't merge this, so we need better testcase).
The point was that with identical .text there doesn't have to be 1:1 mapping
between location_t's (and those now include both file:line:column and BLOCK),
in the above example supposedly there is many:1 mapping from f2 to f1
location_t, but generally there could be overlaps etc.