Bug in GCC ?

Bug in GCC ?

[Sébastien]

He is the sample code I'm trying to compile:

with Ada.Containers.Hashed_Maps;
with Ada.Strings.Unbounded; use Ada.Strings.Unbounded;
with Ada.Strings.Unbounded.Hash;

package SCMAL.Tools.HashedMaps is

     package HashedStrStr is
         new Ada.Containers.Hashed_Maps(
         Key_Type => Unbounded_String,
         Element_Type => Unbounded_String,
         Hash => Ada.Strings.Unbounded.Hash,
         Equivalent_Keys => "=");

end SCMAL.Tools.HashedMaps;


At compile time:

$ gnatmake -Pgentest -cargs -gnatn
gcc-4.1 -c -gnat05 -g3 -O2 -gnatn -I- -gnatA 
/var/local/mscm/ada/tests/generic/scmal-tools-hashedmaps.ads
+===========================GNAT BUG DETECTED==============================+
| 4.1.3 20070929 (prerelease) (Ubuntu 4.1.2-16ubuntu3) 
(i486-pc-linux-gnu) GCC error:|
| in referenced_var_lookup, at tree-dfa.c:578                              |
| No source file position information available                            |
| Please submit a bug report; see http://gcc.gnu.org/bugs.html.            |
| Use a subject line meaningful to you and us to track the bug.            |
| Include the entire contents of this bug box in the report.               |
| Include the exact gcc-4.1 or gnatmake command that you entered.          |
| Also include sources listed below in gnatchop format                     |
| (concatenated together with no headers between files).                   |
+==========================================================================+

It works fine if I remove the -gnatn and -O2

Sebastien

Re: Bug in GCC ?

[Matthew Heaney]

On May 19, 11:50 am, Sébastien <seb.mor...@gmail.com> wrote:
>
> with Ada.Containers.Hashed_Maps;

Why didn't you use the indefinite form?  That would allow you to ditch
the Unbounded_String.

with Ada.Containers.Indefinite_Hashed_Maps;
pragma Elaborate_All (Ada.Containers.Indefinite_Hashed_Maps);

package SCMAL.Tools.HashedMaps is

     package HashedStrStr is
         new Ada.Containers.Indefinite_Hashed_Maps(
         Key_Type => String,
         Element_Type => String,
         Hash => Ada.Strings.Hash,
         Equivalent_Keys => "=");

end SCMAL.Tools.HashedMaps;

Re: Bug in GCC ?

[Sébastien]

> Why didn't you use the indefinite form?  That would allow you to ditch
> the Unbounded_String.


Because I didn't know it ;-)

Re: Bug in GCC ?

[Samuel Tardieu]

Note that GCC 4.1 is quite old, newer GCC do not have this bug.

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/

Re: Bug in GCC ?

[Ludovic Brenta]

Sébastien <seb.morand@gmail.com> writes:
> $ gnatmake -Pgentest -cargs -gnatn
> gcc-4.1 -c -gnat05 -g3 -O2 -gnatn -I- -gnatA
> /var/local/mscm/ada/tests/generic/scmal-tools-hashedmaps.ads
> +===========================GNAT BUG DETECTED==============================+
> | 4.1.3 20070929 (prerelease) (Ubuntu 4.1.2-16ubuntu3)
> (i486-pc-linux-gnu) GCC error:|
> | in referenced_var_lookup, at tree-dfa.c:578                              |
[...]
> It works fine if I remove the -gnatn and -O2

This is a known bug: http://gcc.gnu.org/PR22280

It is fixed in 4.2 and later.

-- 
Ludovic Brenta.

Re: Bug in GCC ?

[Sebastien Morand]

> This is a known bug: http://gcc.gnu.org/PR22280
> 
> It is fixed in 4.2 and later.

Ok that's why I didn't have it on my FreeBSD (gcc 4.3).

So I have to install gcc 4.3 on the linux box too, but it's not packaged, so 
it's boring ...

Re: Bug in GCC ?

[Ludovic Brenta]

Sebastien Morand writes:
>> This is a known bug: http://gcc.gnu.org/PR22280
>>
>> It is fixed in 4.2 and later.
>
> Ok that's why I didn't have it on my FreeBSD (gcc 4.3).
>
> So I have to install gcc 4.3 on the linux box too, but it's not
> packaged, so it's boring ...

It is packaged ad is the default in Debian testing.

-- 
Ludovic Brenta.