[Bug c++/11856] New: Wrong warning regarding assert

[Bug c++/11856] New: Wrong warning regarding assert

[o dot kullmann at swansea dot ac dot uk]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856

           Summary: Wrong warning regarding assert
           Product: gcc
           Version: 3.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: o dot kullmann at swansea dot ac dot uk
                CC: gcc-bugs at gcc dot gnu dot org

<file AssertWarning.cpp> 
---------------------------------------- 
#include <cassert> 
void f(unsigned char c) { 
  assert(0 <= c and c <= 2); 
} 
int main() { 
  f(5); 
} 
---------------------------------------- 
yields: 
> g++ AssertWarning.cpp 
AssertWarning.cpp: In function `void f(unsigned char)': 
AssertWarning.cpp:3: warning: comparison is always true due to limited range 
of data type 
 
which is wrong: Running it yields (as it should be): 
> ./a.out 
a.out: AssertWarning.cpp:3: void f(unsigned char): Assertion `0 <= c and c <= 
2' failed. 
Aborted. 
 
Tested Versions: 3.3 and 3.2.1 (same behaviour). 
(Platform: Linux (Suse 8.0)) 
 
I'm writing a generic library, and this warning shows up all over the 
place, and since the warning messages typically are *very* long (due 
to the deeply nested template names), the output of g++ is rendered 
unreadable. 
 
Since it doesn't seem to be possible to inhibit this warning message (!), 
the only work around is to use 
 
g++ -w 
 
inhibiting *all* warnings.

[Bug c++/11856] Wrong warning regarding assert

[marcus at jet dot franken dot de]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856



------- Additional Comments From marcus at jet dot franken dot de  2003-08-08 14:36 -------
c<=0 is always true for unsigned characters, thats where the warning comes 
from. 
 
your code reduces to: 
assert(c <= 2);

[Bug c++/11856] Wrong warning regarding assert

[o dot kullmann at swansea dot ac dot uk]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856



------- Additional Comments From o dot kullmann at swansea dot ac dot uk  2003-08-08 15:17 -------
Subject: Re:  Wrong warning regarding assert

On Fri, Aug 08, 2003 at 02:36:08PM -0000, marcus at jet dot franken dot de wrote:
> Date: 8 Aug 2003 14:36:08 -0000
> From: "marcus at jet dot franken dot de" <gcc-bugzilla@gcc.gnu.org>
> To: o.kullmann@swansea.ac.uk
> In-Reply-To: <20030808142556.11856.o.kullmann@swansea.ac.uk>
> Reply-To: gcc-bugzilla@gcc.gnu.org
> X-Bugzilla-Reason: Reporter
> X-SA-Exim-Mail-From: dberlin@gcc.gnu.org
> Subject: [Bug c++/11856] Wrong warning regarding assert
> X-Spam-Status: No, hits=-101.5 required=5.5
> 	tests=AWL,BAYES_30,IN_REP_TO,REFERENCES,USER_IN_WHITELIST
> 	version=2.55
> X-Spam-Level: 
> X-Spam-Checker-Version: SpamAssassin 2.55 (1.174.2.19-2003-05-19-exp)
> X-SA-Exim-Version: 3.0 (built Sat May 17 23:48:01 PDT 2003)
> X-UIDL: !lB"!9IP"!iVl"!f/!"!
> 
> PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.
> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856
> 
> 
> 
> ------- Additional Comments From marcus at jet dot franken dot de  2003-08-08 14:36 -------
> c>=0 is always true for unsigned characters, thats where the warning comes 
> from. 
>  
> your code reduces to: 
> assert(c <= 2); 
> 
>

I see; but the problem is that with generic code, where function f in fact is
a function template, the warning becomes a bad thing: The assertion c>=0
in general is necessary (for signed cases), but for all unsigned cases
we get a warning message (as I said, the typical monsters, pages and pages for
just one message).

My general observation is, that the warning system of g++ doesn't fit well with
generic programming, where all those trivial situations (which triggers a warning)
in fact often are essential.

So perhaps the solution is to inhibit all warnings. But intuitively this seems too
bold to me. I checked again the documentation of gcc/g++, but I couldn't find anything
to turn off the above assert warning. Perhaps for generic programming it would be a good
thing to be able to switch off all "triviality warnings".

Oliver

 
> 
> ------- You are receiving this mail because: -------
> You reported the bug, or are watching the reporter.
>

[Bug c++/11856] Wrong warning regarding assert

[o dot kullmann at swansea dot ac dot uk]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856



------- Additional Comments From o dot kullmann at swansea dot ac dot uk  2003-08-08 15:29 -------
Subject: Re:  Wrong warning regarding assert (addition)

> > ------- Additional Comments From marcus at jet dot franken dot de  2003-08-08 14:36 -------
> > c>=0 is always true for unsigned characters, thats where the warning comes 
> > from. 
> >  

I'm sorry, but I missed the point:

In the gcc-info-documentation I have it says under option "-W":

`-W'
      Print extra warning messages for these events:

* An unsigned value is compared against zero with `<' or `<='.


So I think the error here is that I get the *extra* warnings of "-W" without
specifying "-W".

Oliver

[Bug c++/11856] Wrong warning regarding assert

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-08 16:10 -------
With this code:
template <typename t>
void f(t c) {
  assert(0 <= c and c <= 2);
}
int main() {
  f<unsigned char>(5);
}
I only get a warning on instantiation.

[Bug c++/11856] Wrong warning regarding assert

[o dot kullmann at swansea dot ac dot uk]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856



------- Additional Comments From o dot kullmann at swansea dot ac dot uk  2003-08-08 16:28 -------
Subject: Re:  Wrong warning regarding assert [Improved example]

> pinskia at gcc dot gnu dot org changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>            Keywords|                            |diagnostic
> 
> 
> ------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-08 16:10 -------
> With this code:
> template <typename t>
> void f(t c) {
>   assert(0 <= c and c <= 2);
> }
> int main() {
>   f<unsigned char>(5);
> }
> I only get a warning on instantiation.
> 
>

I see, I oversimplified the problem a bit. Here is a better example:
------------------------------
#include <cassert>
template <typename Int, Int D>
void f(Int x) {
  assert(0 <= x and x <= D);
}
int main() {
  f<unsigned char, 2>(5);
}
-----------------------------
> g++ AssertWarning.cpp
AssertWarning.cpp: In function `void f(Int) [with Int = unsigned char, Int D =
   2]':
AssertWarning.cpp:7:   instantiated from here
AssertWarning.cpp:4: warning: comparison is always true due to limited range of
   data type
> ./a.out
a.out: AssertWarning.cpp:4: void f(Int) [with Int = unsigned char, Int D = 2]: Assertion `0 <= x and x <= D' failed.
Aborted.


Oliver

[Bug c++/11856] unsigned warning in template

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2003-08-10 02:35:07
               date|                            |
            Summary|Wrong warning regarding     |unsigned warning in template
                   |assert                      |


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-10 02:35 -------
I can confirm this on the mainline (20030809).  I like this warning but this case looks 
maybe too over warning but the warning is still true (usigned <= 0 is always true).

[Bug c++/11856] unsigned warning in template

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856



------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-10 02:36 -------
I really meant unsigned >= 0 is always true.

[Bug c++/11856] unsigned warning in template

[dhazeghi at yahoo dot com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856


dhazeghi at yahoo dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.4                         |---

[Bug c++/11856] unsigned warning in template

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-05-29 14:49 -------
*** Bug 15723 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |igodard at pacbell dot net


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11856