Re: oh. the sizeof() problem is worse than I thougt
On Thu, 1 Mar 2001, Matt Wilson wrote:
> On Thu, Mar 01, 2001 at 03:43:15PM -0500, Stuart Anderson wrote:
> >
> > Yes, the Type stuff should be architecture sensitive. It is designed to be,
> > but since not everything in that areas has been 100% finished, it's possible
> > some changes are still needed. Also, the data that has been loaded is IA32
> > in origin. I need to get comparable data from other architectures, and sift
> > through the differences.
>
> OK, here's the output of a run on Alpha...
>
> Your checkin today seems to be missing X11_X_h.c and X11_Xutil_h.c.
Sure enough. I'll fix that.
> BTW - elfchk seems to do only 32 bit elf?
Correct. The bulk of the code was actually written pre-64 bit. Making a pass
over it is on the list.
> sizeof(GLXFBConfig) is 8 instead of 4
> UPDATE Type SET Tsize=8 WHERE Tname='GLXFBConfig';
Everything derived from a fundamental type will be affected. I need to build
& import the header data on a 64 bit system, as a first broad pass at it,
and then resolve what is left over then. An alternate method would be to
adapt this output to insert new records with a different architecture,
and make sure the mktests script is architecture sensitive.
Stuart
Stuart R. Anderson anderson@metrolink.com
Metro Link Incorporated South Carolina Office
5807 North Andrews Way 129 Secret Cove Drive
Fort Lauderdale, Florida 33309 Lexington, SC 29072
voice: 954.660.2500 voice: 803.951.3630
http://www.metrolink.com/ XFree86 Core Team
Creative Applications Lab Chair - SIGGRAPH 2001
Reply to: