Re: Win32 Integral size mismatch warnings.

From: d. hall (dhall@OOI.NET)
Date: 08/24/98


>>>>>> thus on Mon, 24 Aug 1998 15:45:49 -0400, George wrote:

> Structs.h has to be included before utils.h, and always is.  I have a
> list of all the integral mismatch warnings from someone with Windows and
> at least the first three (all I checked) coincided with GET_ROOM_VNUM.
> Hm...

I understand this, as this has always been for stock circle since 3.x first
went beta, and 2.0 before that, etc... ad nauseum, but; the more correct
method(tm) is to:

/* prevent multiple includes */
#ifndef __CIRCLE_STRUCTS_H__
#define __CIRCLE_STRUCTS_H__
/* structs.h file */
#endif __CIRCLE_STRUCTS_H__

And including structs.h within utils.h when utils.h requires typecast from
structs.h, for the sake of ease of use for other programmers.

Anyways, back to the subject, digging around, I found;

Although room_vnum is typedef'ed from sh_int (which is further typedef'ed
from signed short int), how well does the Win32 compiler deal with implicit
3rd generation typedef's.  The only value that is explicitly cast as
room_vnum is load_room(player_special_data_saved) and number(room_data).

real_atrium in hcontrol_build_house() is sh_int (which is technically the
same as room_vnum), but some compilers are extremely fussy.  SGI's native
compiler has levels of complaints that far exceed gcc in -Wall mode.

Could someone send me the complaints output?

d.


     +------------------------------------------------------------+
     | Ensure that you have read the CircleMUD Mailing List FAQ:  |
     | http://democracy.queensu.ca/~fletcher/Circle/list-faq.html |
     +------------------------------------------------------------+



This archive was generated by hypermail 2b30 : 12/15/00 PST