[Bug c/30368] New: wrong result

[Bug c/30368] New: wrong result

[lidaobing at gmail dot com]

for the following example[1], it should exit with 0, but under gcc-4.0 gcc-4.1
gcc-4.2 and all optimize level(-O0, -O1, -O2, -O3), it exit with 1;

check gcc versions in [2]

[1]
/*  begin */
struct A { int b; };

int main(int argc, char **argv)
{
    struct A* a = 0;
    if (&a->b) return 1;
    return 0;
}
/* end */

[2] 
gcc version 4.0.4 20060904 (prerelease) (Debian 4.0.3-7)
gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)
gcc version 4.2.0 20061003 (experimental) (Debian 4.2-20061003-1)


-- 
           Summary: wrong result
           Product: gcc
           Version: 4.0.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: lidaobing at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368

[Bug c/30368] wrong result

[lidaobing at gmail dot com]

------- Comment #1 from lidaobing at gmail dot com  2007-01-04 14:25 -------
orignal reported by RoachCock in newsmth.net BBS


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368

[Bug c/30368] wrong result

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2007-01-05 05:23 -------
Actually this is undefined by the C standard as you are deferencing a null
pointer, yes &a->b is implemented as a pointer arthematic but it is still a
deferencing according to the C standard.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368

[Bug c/30368] wrong result

[pinskia at gcc dot gnu dot org]

------- Comment #3 from pinskia at gcc dot gnu dot org  2007-01-05 20:13 -------
> anther, consider an example definite[2] of 'offsetof', if you think
> that is undefined, it's almost impossible to give a definite of
> offsetof.

> #define offsetof(TYPE,MEMBER)   ((size_t)&((TYPE*)0)->MEMBER)

The C standard still says that is undefined.  See 6.5.3.2/4.
Also GCC has a builtin for offsetof to get around the undefinedness of the
above.

You can also define offsetof like:
#define offsetof(TYPE,MEMBER)   (((size_t)&((TYPE*)128)->MEMBER) - 128)

Which gets around the undefined issue.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368

[Bug c/30368] wrong result

[gdr at integrable-solutions dot net]

------- Comment #4 from gdr at integrable-solutions dot net  2007-01-05 21:11 -------
Subject: Re:  wrong result

"pinskia at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org> writes:

| > anther, consider an example definite[2] of 'offsetof', if you think
| > that is undefined, it's almost impossible to give a definite of
| > offsetof.
| 
| > #define offsetof(TYPE,MEMBER)   ((size_t)&((TYPE*)0)->MEMBER)
| 
| The C standard still says that is undefined.  See 6.5.3.2/4.
| Also GCC has a builtin for offsetof to get around the undefinedness of the
| above.

The C front-end should probably warn -- the C++ front-end does.


-- Gaby


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368

[Bug c/30368] wrong result

[manu at gcc dot gnu dot org]

------- Comment #5 from manu at gcc dot gnu dot org  2007-01-26 23:14 -------
(In reply to comment #4)
> Subject: Re:  wrong result
> 
> "pinskia at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org> writes:
> 
> | > anther, consider an example definite[2] of 'offsetof', if you think
> | > that is undefined, it's almost impossible to give a definite of
> | > offsetof.
> | 
> | > #define offsetof(TYPE,MEMBER)   ((size_t)&((TYPE*)0)->MEMBER)
> | 
> | The C standard still says that is undefined.  See 6.5.3.2/4.
> | Also GCC has a builtin for offsetof to get around the undefinedness of the
> | above.
> 
> The C front-end should probably warn -- the C++ front-end does.

Does it? With which options? I wasn't able to get a warning or error.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368

[Bug c/30368] missing warning for dereferencing null pointer

[manu at gcc dot gnu dot org]

------- Comment #6 from manu at gcc dot gnu dot org  2007-03-13 16:12 -------
So the conclusion is that we should emit a warning for the undefinedness.


-- 

manu at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2007-03-13 16:12:46
               date|                            |
            Summary|wrong result                |missing warning for
                   |                            |dereferencing null pointer


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368

[Bug c/30368] missing warning for dereferencing null pointer

[manu at gcc dot gnu dot org]

------- Comment #7 from manu at gcc dot gnu dot org  2007-03-13 16:22 -------
and as such, this is a duplicate of bug 16351.

*** This bug has been marked as a duplicate of 16351 ***


-- 

manu at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30368