On Sun, 23 Dec 2001, Christoph Reichenbach wrote:

> > #291 (PQ2 shopwoman bug) still appears to be broken. It looks like
> > FreeSCI's values for brRect and nsRect are totally wrong (actually, only
> > the left/right values for those rects).
>
> Do you have those available somewhere? I can't really test very well from
> here.

I have updated the screenshots on my website from the SSCI debugger. I
have noticed a pattern: the horizontal coordinates (x, brLeft, brRight,
etc) for both haines and shopwoman seem to be offset by -50 or so. I
believe that because of this, the MoveTo gets very confused and never
completes. See the screen shots and the FreeSCI debugger output @
http://www.clock.org/~matt/freesci .


> > In SSCI, shopwoman has a signal of 0 and in FreeSCI, it is 0x0400 (1024).
>
> That would indicate that she has collided with something. Could you test
> this with +b so that we know what she hit, please?
>
> (Check for 'finished mover by collision'; the cause can usually be derived
> from the text printed before that.)

That message never happens, it appears to go into an infinite loop. The
debug output is in the pq2-inspect.txt file mentioned above.


> > I can't verify any more of PQ2 until the shopwoman bug
> > is fixed, of course.
>
> This is the only case of the brRect being off that I know of, and it's
> quite surprising. This used to work earlier, right? We should check if the
> brRect used to be correct.

I only have recent CVS and 0.3.2a on my machine, getting a whole source
tree from a previous date would take awhile on this dialup. 0.3.2a crashes
for me pretty immediately in PQ2, so I can't give any further information
on this until I get back (on the 27th). I'm assuming this bug did not
exist before, since PQ2 was beaten by two different people before.
Hmm.. perhaps doing a CVS binary search between the time it was beaten and
now would be a good idea..


> Only the sound server remains to be fixed now.

I'm working on this (from win32 side). I'll let the list know if/when I
get somewhere.


--
http://www.clock.org/~matt


Reply via email to