• d-bugmail puremagic.com (27/27) Mar 10 2006 http://d.puremagic.com/bugzilla/show_bug.cgi?id=33
• d-bugmail puremagic.com (18/18) Mar 13 2006 http://d.puremagic.com/bugzilla/show_bug.cgi?id=33
• Thomas Kuehne (15/23) Mar 21 2006 -----BEGIN PGP SIGNED MESSAGE-----
• d-bugmail puremagic.com (9/9) Apr 28 2006 http://d.puremagic.com/bugzilla/show_bug.cgi?id=33

d-bugmail puremagic.com writes:

http://d.puremagic.com/bugzilla/show_bug.cgi?id=33

           Summary: No short-circuit evaluation for compile-time
                    expressions?
           Product: D
           Version: 0.149
          Platform: PC
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: bugzilla digitalmars.com
        ReportedBy: sean f4.ca


I think this is arguably not a bug in the same way that compile-time "?:"
expressions evaluate both their arguments regardless of which is selected, but
it's definately non-intuitive:


C:\code\d\bugs>type 149_2.d
import std.c.stdio;

void main()
{
    static if( int.mangleof.length > 1 && int.mangleof[1] == 'x' )
        printf( "'x' as second char\n" );
}
C:\code\d\bugs>dmd 149_2.d
149_2.d(5): array index [1] is outside array bounds [0 .. 1]

C:\code\d\bugs>


-- 

d-bugmail puremagic.com writes:

http://d.puremagic.com/bugzilla/show_bug.cgi?id=33






Definitely a bug -- but maybe not in &&. This example compiles:

void main()
{
    const char [] s = int.mangleof;
    static if( s.length > 1 && s[1] == 'x' )
        printf( "'x' as second char\n" );
}

And this one gives a quite bizarre error message:

void main()
{
    const char [] s = (int*).mangleof;
    static if( s.length > 1 && s[2] == 'x' )
        printf( "'x' as second char\n" );
}

something about TOK58 doesn't evaluate to a boolean.


-- 

Thomas Kuehne <thomas-dloop kuehne.cn> writes:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

d-bugmail puremagic.com schrieb am 2006-03-13:
 And this one gives a quite bizarre error message:

 void main()
 {
     const char [] s = (int*).mangleof;
     static if( s.length > 1 && s[2] == 'x' )
         printf( "'x' as second char\n" );
 }

 something about TOK58 doesn't evaluate to a boolean.

The code above is buggy. s="ab" will result in an  ArrayBoundsError.

Added to DStress
http://dstress.kuehne.cn/run/m/mangleof_15_A.d
http://dstress.kuehne.cn/run/m/mangleof_15_B.d
http://dstress.kuehne.cn/run/m/mangleof_15_C.d
http://dstress.kuehne.cn/run/m/mangleof_15_D.d

Thomas


-----BEGIN PGP SIGNATURE-----

iD8DBQFEH+AF3w+/yD4P9tIRAic1AKCSrOmy1GOb0Lsry6wfLQ2H7b/IFQCgvDT9
BPkzHqg8uz8ZkR+y3x+WD3g=
=dpwQ
-----END PGP SIGNATURE-----

d-bugmail puremagic.com writes:

http://d.puremagic.com/bugzilla/show_bug.cgi?id=33


bugzilla digitalmars.com changed:

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





Fixed 0.155


--