DIB_DirectDrawSurface_BltFast signed/unsigned comparision bug

[Piotr Pawlow]

Hello,

I spotted this bug while trying to fix mouse pointer trails problem in 
Warlords Battlecry. All widths, heights and left/top coordinates are of DWORD 
type, so substracting them from each other always give positive numbers. Thus 
h and w are always >= 0, so DDERR_INVALIDRECT is returned only in a couple of 
boundary cases.

I just added explicit casting to int where needed. It didn't fix the mouse 
trails, but previously missing mouseover effect on icons at the main screen 
appeared, so I guess it is a change in the right direction :)

Patch attached.

[Lionel Ulmer]

> I just added explicit casting to int where needed. It didn't fix the mouse 
> trails, but previously missing mouseover effect on icons at the main screen 
> appeared, so I guess it is a change in the right direction :)

While your patch is certainly OK, if 'ddesc.dwWidth-dstx' is lower than
zero, it means that there is nothing to do (as the destination X coordinate
is out of the screen). Same for the source rectangle.

So could you try rewriting your patch with something like that :

if (ddesc.dwWidth<=dstx) return DD_OK;
(and same for Y and srcs).

And check if it gives you the same result as your previous patch. Maybe it's
some added code, but well, it will be a bit cleaner IMHO.

Anyway, thanks for fixing bugs in the DDraw code, we certainly needs more
contributors :-)

          Lionel

-- 
		 Lionel Ulmer - http://www.bbrox.org/

[Piotr Pawłow]

On Monday 26 January 2004 22:27, Lionel Ulmer wrote:
> While your patch is certainly OK, if 'ddesc.dwWidth-dstx' is lower than
> zero, it means that there is nothing to do (as the destination X coordinate
> is out of the screen). Same for the source rectangle.

Are you sure about this? Isn't INVALIDRECT returned in this case? I am asking 
these questions, because I could not find any documentation in MSDN 
explaining when is INVALIDRECT returned. Do you have any test results? Maybe 
I should write a test?
> And check if it gives you the same result as your previous patch. Maybe
> it's some added code, but well, it will be a bit cleaner IMHO.
> Anyway, thanks for fixing bugs in the DDraw code, we certainly needs more
> contributors :-)

Thanks, I will look into this as soon as I find some free time. Currently I 
have more important work to do for a living, so Wine has to wait :/

-- 
 Piotr Pawłow
 mailto:pp(a)siedziba.pl

[Lionel Ulmer]

> Are you sure about this? Isn't INVALIDRECT returned in this case? I am asking 
> these questions, because I could not find any documentation in MSDN 
> explaining when is INVALIDRECT returned. Do you have any test results? Maybe 
> I should write a test?

Well, MSDN usually does not document at all the border / failure cases... So
best is to write a test program.

If you do not have the mean to do it, I may do it later this week (or on the
plane on Thursday :-) ).

    Lionel

-- 
		 Lionel Ulmer - http://www.bbrox.org/

[Piotr Pawłow]

On Wednesday 04 February 2004 23:15, Lionel Ulmer wrote:
> If you do not have the mean to do it, I may do it later this week (or on
> the plane on Thursday :-) ).

I will be busy at least till Saturday, so if you would like to do it on a 
boring flight, go ahead :)

-- 
 Piotr Pawłow
 mailto:pp(a)siedziba.pl