On Thu, 29 Nov 2001, Christoph Reichenbach wrote: > > Where do I go from here for further debugging? > > First, let me formulate a theory for the CB1 bug, please (can't think of > any relevance of this change wrt the PQ2 one): > Since this changed the behaviour of dynview collissions, it appears likely > that Laura collides with one of the two guys who just moved inside the > house. This can be checked by turning on the debuglog for bresenham, > watching what Laura collides with, and getting the object's name, getting > its position and flags and, if the theory is correct so far, comparing > these values to Sierra SCI's values in the same scene (not waiting for > Laura to halt, of course- the obstacle's values should remain constant as > soon as its mover has finished, which should happen even before Laura > comes near the door).
I'm not exactly sure what to make of the debug output, but here goes... The last collision check is: " kernel: (E:\src\freesci\src\engine\kgraphics.c L738): Checking collision: (150, 129) to (17,2) ([150..167]x[119..121]), obj=0684, sig=6c02, cliplist=7924 kernel: (E:\src\freesci\src\engine\kgraphics.c L747): edgehit = 0012 kernel: (E:\src\freesci\src\engine\kgraphics.c L760): Checking vs dynviews: kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=008a, z=0080 obj=9ca4 kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=009a, z=0080 obj=8f46 kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=009a, z=0080 obj=8f94 kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=0002, z=0000 obj=8ee2 kernel: (E:\src\freesci\src\engine\kgraphics.c L697): against (181,116) to (2 00,118) kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=0002, z=0000 obj=8e7e kernel: (E:\src\freesci\src\engine\kgraphics.c L697): against (167,116) to (1 88,120) kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=4010, z=4000 obj=8ca0 kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=4002, z=4000 obj=8cf2 kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L691): OtherSignal=4002, z=4000 obj=8d40 kernel: (E:\src\freesci\src\engine\kgraphics.c L710): (no) kernel: (E:\src\freesci\src\engine\kgraphics.c L772): -> 4000" heapobj on 0x8d40 shows: "Object Splash2 Species=2222, Superclass=2222 Local variables @ 0x90b6 Variable selectors: species[0000] = 2222 %Prop superClass[0001] = 2222 %Prop -info-[0002] = 0000 (0) name[0017] = 9063 Warning: Unaligned unsigned read from 9063 y[0003] = 00b6 (182) x[0004] = 00b7 (183) z[0055] = 0000 (0) heading[003a] = 0000 (0) yStep[0037] = 0002 (2) view[0005] = 0001 (1) loop[0006] = 0006 (6) cel[0007] = 0006 (6) priority[003f] = 000e (14) underBits[0008] = 0000 (0) signal[0011] = 4002 nsTop[0009] = 00b6 (182) nsLeft[000a] = 00b7 (183) nsBottom[000b] = 00b7 (183) nsRight[000c] = 00b8 (184) lsTop[000d] = 0000 (0) lsLeft[000e] = 0000 (0) lsBottom[000f] = 0000 (0) lsRight[0010] = 0000 (0) brTop[0013] = 00b5 (181) brLeft[0014] = 00b0 (176) brBottom[0015] = 00b7 (183) brRight[0016] = 00bf (191) cycleSpeed[00a4] = 0000 (0) script[0076] = 0000 (0) cycler[00a5] = 0000 (0) timer[0071] = 0000 (0) -- View information: cel 1/6/6 at (183,182,0) Priority = 14 (band starts at 180) nsRect: [183..184]x[182..183] lsRect: [0..0]x[0..0] brRect: [176..191]x[181..183] new nsRect: [183..184]x[182..183] new clipped nsRect: [183..184]x[182..183] new brRect: [183..184]x[181..183] signals = 4002: 0002: updated 4000: ignore_actor" There are differences in this object (Splash2) between FreeSCI and Sierra SCI. The different values are (var: Original SCI, Freesci): lsTop: 182, 0 lsLeft: 183, 0 lsBottom: 183, 0 lsRight: 184, 0 So, the values for lsRect in FreeSCI are zero'd for some reason. Once again, I ask the question.. where do I go from here? Should I try and verify the PQ2 bug in the same way, or is this enough information to formulate a fix? Thanks for the help, I'm learning a great deal in this exercise :) -- http://www.clock.org/~matt
