http://bugs.winehq.org/show_bug.cgi?id=50876
ttoocs@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ttoocs@gmail.com
--- Comment #12 from ttoocs@gmail.com --- Was curious to play simgolf myself and ran into this. So ran simgolf with WINEDBG=+all, and traced it down a trace:opengl:wglChoosePixelFormat with PDF_DRAW_BITMAP. Which seemed to match this and to the same functions discussed in https://bugs.winehq.org/show_bug.cgi?id=58662 . There's a commit in the wglChoosePixelFormat function that changes how it handles that flag, prior it would ignore it (I think this is how we get the trippy colors) and now fails and gives me just black terrain. I believe bug 58662 is this, where no valid pixel formats existing of r5g5b5a1... as I guess from my searching only some windows opengl 1.1 software rendering driver actually supported PDF_DRAW_BITMAP anyhow.
It seems Zowie is working on a solution on bug 58662, and while a dev myself, he already started on it a week ago, and presumably knows something of the codebase.
(I accidentally lost the original 1.5Gb logfile, but even in the "short" 400Mb one, I was doing line-jumps of 50,000k-500,000k to find when functions actually seemed to change. grep -n "trace:bidi:BIDI_Reorder" was a great bookmark/chapters for simgolf, "... one moment ..." and then the next is seems to be related the error. There's also a CALL Terrain.?initSystem@Terrain<garbled> which occurs right before calls to gdu32 and opengl32 start, tho maybe prior ones as well (lost the log to check) )