Re: wine/ programs/rundll32/rundll32.c dlls/kernel ...

[Martin Wilck]

Am Fre, 2002-11-15 um 05.17 schrieb Alexandre Julliard:
> ChangeSet ID:	6302
> Modified files:
>    programs/rundll32: rundll32.c 
>    dlls/kernel    : computername.c
>
> Log message:
>    Warning fixes.

Your compiler seems to issue more warnings than mine :-/
It is good to see these patches are really reviewed. The lenW thingy
is really embarrassing for me.

Two questions/remarks:

@@ -426,12 +426,13 @@
                 memcpy( name, buf, len );
                 name[len] = 0;
                 *size = len;
+                ret = TRUE;

This shouldn't be necessary because the whole block is executed only
if ret == TRUE.

}
         }
         __EXCEPT(page_fault)
         {
             SetLastError( ERROR_INVALID_PARAMETER );
-            ret = FALSE;
+            return FALSE;
         }
         __ENDTRY

What is the point of this one? (applies also to the last hunk)

Martin

wine/ programs/rundll32/rundll32.c dlls/kernel ...

[Alexandre Julliard]

Martin Wilck  writes:
> @@ -426,12 +426,13 @@
>                  memcpy( name, buf, len );
>                  name[len] = 0;
>                  *size = len;
> -
>            ret = TRUE;
>
> This shouldn't be necessary because the whole block is executed only
> if ret == TRUE.

I know that, but the compiler doesn't... Because the exception macros
use setjmp the compiler has to assume that variables change, even if
they don't. It could be possible to restructure the code to avoid the
extra assignment, I just didn't want to risk breaking it.
>      {
>          SetLastError( ERROR_INVALID_PARAMETER );
> -
>        ret = FALSE;
> -
>        return FALSE;
>    }
>    __ENDTRY
>
> What is the point of this one? (applies also to the last hunk)

It makes no difference to the code, it's only to help with the
warnings.