[Bug gdb/24318] GDB prints the wrong value

[Bug gdb/24318] GDB prints the wrong value

[ssbssa at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=24318

Hannes Domani <ssbssa at sourceware dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ssbssa at sourceware dot org

--- Comment #1 from Hannes Domani <ssbssa at sourceware dot org> ---
(In reply to e from comment #0)
> main.c :
> uint64_t * tab;
> 
> test.c :
> extern uint64_t tab[];

This looks like undefined behavior to me.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/24318] GDB prints the wrong value

[jbrejon at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=24318

--- Comment #2 from e <jbrejon at gmail dot com> ---
(In reply to Hannes Domani from comment #1)
> (In reply to e from comment #0)
> > main.c :
> > uint64_t * tab;
> > 
> > test.c :
> > extern uint64_t tab[];
> 
> This looks like undefined behavior to me.

I believe the C standard is not the issue here.
Its about gdb consistency.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/24318] GDB prints the wrong value

[ssbssa at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=24318

--- Comment #3 from Hannes Domani <ssbssa at sourceware dot org> ---
(In reply to e from comment #2)
> (In reply to Hannes Domani from comment #1)
> > (In reply to e from comment #0)
> > > main.c :
> > > uint64_t * tab;
> > > 
> > > test.c :
> > > extern uint64_t tab[];
> > 
> > This looks like undefined behavior to me.
> 
> I believe the C standard is not the issue here.
> Its about gdb consistency.

How could there be consistency when tab isn't consistent the same type
everywhere?

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/24318] GDB prints the wrong value

[jbrejon at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=24318

--- Comment #4 from e <jbrejon at gmail dot com> ---
(In reply to Hannes Domani from comment #3)
> (In reply to e from comment #2)
> > (In reply to Hannes Domani from comment #1)
> > > (In reply to e from comment #0)
> > > > main.c :
> > > > uint64_t * tab;
> > > > 
> > > > test.c :
> > > > extern uint64_t tab[];
> > > 
> > > This looks like undefined behavior to me.
> > 
> > I believe the C standard is not the issue here.
> > Its about gdb consistency.
> 
> How could there be consistency when tab isn't consistent the same type
> everywhere?

My original post contains two experiments in which the same exact code is
executed. In those experiments GDB does not prints the same values for the same
elements at the same moment.
Same code executed (consistentcy), not the same output (unconsistency).

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/24318] GDB prints the wrong value

[tromey at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=24318

Tom Tromey <tromey at sourceware dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
                 CC|                            |tromey at sourceware dot org
         Resolution|---                         |INVALID

--- Comment #5 from Tom Tromey <tromey at sourceware dot org> ---
The program here is invalid, as Hannes pointed out.
Now, that's exactly when you'd want to use gdb.
However, in this case, gdb ignores the declaration in test.c
in favor of the definition of the object.
IOW, it is doing the right thing.
Expecting it to mimic the undefined behavior of the source
seems a bit weird.

Anyway I don't think we should try to change gdb to do this.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/24318] GDB prints the wrong value

[jbrejon at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=24318

--- Comment #6 from e <jbrejon at gmail dot com> ---
(In reply to Tom Tromey from comment #5)
> The program here is invalid, as Hannes pointed out.
it is.


> However, in this case, gdb ignores the declaration in test.c
> in favor of the definition of the object.
> IOW, it is doing the right thing.
Not sure what you mean by that as GDB doesn't prints the same thing in the two
execution. I don't think it favor anything as its not consistent.
Again:
GDB first instance:
...
test () at test.c:8
8               printf("tab[0] == %"PRIx64"\n", tab[0]);
(gdb) p tab[0] <--------------------------------------------(b)
...
GDB second instance:
...
test () at test.c:8
8               printf("tab[0] == %"PRIx64"\n", tab[0]);
(gdb) p/x tab[0] <-------------------------------------------(b)
...

What changed? In the first instance we printed (a) first. It looks like some
kind of type caching. I'm not sure if its intended but that was definitly not
what I expected.

-- 
You are receiving this mail because:
You are on the CC list for the bug.