http://d.puremagic.com/issues/show_bug.cgi?id=285

           Summary: Struct method null pointer assert has line number of "0"
           Product: D
           Version: 0.163
          Platform: PC
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: bugzilla@digitalmars.com
        ReportedBy: jarrett.billingsley@gmail.com


struct S
{
        void fork()
        {

        }
}

void main()
{
        S* s = null;
        s.fork();
}

This causes an AssertError with the correct module name, but with a line number of 0, which is really, really unhelpful.

I didn't know DMD even inserted this null check, I think it's pretty cool :) Would be nice if it included the message "struct 'this' pointer is null" or "struct method called through null pointer" or something along those lines.


-- 

d-bugmail@puremagic.com wrote:
> http://d.puremagic.com/issues/show_bug.cgi?id=285
> 
>            Summary: Struct method null pointer assert has line number of "0"
>            Product: D
>            Version: 0.163
>           Platform: PC
>         OS/Version: Windows
>             Status: NEW
>           Severity: normal
>           Priority: P2
>          Component: DMD
>         AssignedTo: bugzilla@digitalmars.com
>         ReportedBy: jarrett.billingsley@gmail.com
> 
> 
> struct S
> {
>         void fork()
>         {
> 
>         }
> }
> 
> void main()
> {
>         S* s = null;
>         s.fork();
> }
> 
> This causes an AssertError with the correct module name, but with a line number
> of 0, which is really, really unhelpful.
> 
> I didn't know DMD even inserted this null check, I think it's pretty cool :) Would be nice if it included the message "struct 'this' pointer is null" or
> "struct method called through null pointer" or something along those lines.
> 
> 


Cute and all; but I sure hope such codegen is disabled via a switch such as -release ...

"kris" <foo@bar.com> wrote in message news:ebln9m$24kh$1@digitaldaemon.com...

> Cute and all; but I sure hope such codegen is disabled via a switch such as -release ...

Indeed it is, just like array bounds checking and the like.  In fact, all it does is insert an implicit "assert(this !is null)", and since it's like any other assert, it compiles out in release mode.

d-bugmail@puremagic.com schrieb am 2006-08-12:
> http://d.puremagic.com/issues/show_bug.cgi?id=285

> struct S
> {
>         void fork()
>         {
>
>         }
> }
>
> void main()
> {
>         S* s = null;
>         s.fork();
> }
>
> This causes an AssertError with the correct module name, but with a line number of 0, which is really, really unhelpful.

Added to DStress as http://dstress.kuehne.cn/norun/s/struct_26_A.d

Thomas

http://d.puremagic.com/issues/show_bug.cgi?id=285


smjg@iname.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |smjg@iname.com
           Keywords|                            |diagnostic




------- Comment #2 from smjg@iname.com  2006-09-09 18:27 -------
It's been suggested before that we should have some kind of error that unambiguously indicates an attempt to dereference a null pointer.  This was mainly in reference to class object references, but I suppose it would apply to structs just as well.


-- 

http://d.puremagic.com/issues/show_bug.cgi?id=285


bugzilla@digitalmars.com changed:

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




------- Comment #3 from bugzilla@digitalmars.com  2006-09-19 15:24 -------
Fixed in DMC 0.167.


--