Hi, > Now that things are stable, I was playing the games a bit more and > noticed: > > In HQ1, in the opening scene, the Goon's yo-yo and the Sherriff's pipe are > not animated as they should be. Can't reproduce. Probably a performance issue; some games turn off animations if they believe that the host is too slow. Does this affect you in 1x1 mode, too? > In KQ4, in the opening credits, the bottom of the last row of credits is > cut off on several of the credit lists. This also happens with text boxes > located near the bottom of the screen during the story introduction. This also seems to affect the LSL3 intro (don't recall seeing it in text boxes, though). > I also saw these in KQ4: > Warning: With view.000: loop 2 > maxloop 1 > Warning: With view.325: loop 3 > maxloop 3 (Should be '>=', of course). Lars: How does Sierra SCI handle this? If DirToLoop() is invoked on an object with too few loops, does it set it to zero (as we do)? Or does it do something completely different? > reported several times, and then: > > Warning: With view.382: loop 3 > maxloop 2 > Warning: Warning: Kernel function 0x55 invoked: NOP This is probably DoAvoider (see 0.3.3 TODO), present in very early KQ4 implementations. > At which point, I tried to control+break out and.. freesci crashed. > Backtrace: > > gfx_tools.c(164): gfx_free_pixmap: > pxm->ID = -1 > pxm->loop = -1 > pxm->cel = -777 > pxm->colors_nr = 14738836 > pxm->colors = NULL > pxm->flags = 0 > pxm->xoffset = 286364986 > pxm->yoffset = 1200 > pxm->index_data = (invalid pointer) > pxm->alpha_map = NULL > pxm->data = NULL > pxm->xl = 17 > pxm->yl = 32 > pxm->internal->handle = 0 > pxm->internal->info = NULL > > sdl_driver.c(313): sdl_exit: > operations.c(602): gfxop_exit: > main.c(1047): main Thanks! I'll have a look at that later. > One last bug, when pic0 is scaled, the volcano in the background during > the intro has a line misdrawn. There is a gap between where the line is > drawn and where the volcano pic is. I assume you mean LSL3. This is true and a good example for why we need the comparisons against unscaled pictures to prevent this from getting any worse. I'm open to suggestions regarding this problem... llap, Christoph
