Perhaps use detours to hook the ChangeDisplaySettingsEx function and dump the devmode structure to a file? Regards, mark
On Wed, Mar 5, 2008 at 9:09 AM, Denver Gingerich denver@ossguy.com wrote:
I am trying to trace the sequence of screen mode (resolution/bpp/refresh rate) changes that occur when StarCraft is run in Windows so that I can make Wine emulate Windows' behavior. Right now I'm trying to see exactly what resolution, bits per pixel, and refresh rate StarCraft is using. Unfortunately, I can't just go into Display Properties while StarCraft is running because it switches the resolution back when you Alt-Tab out of the game.
My current plan is to install an SSH or telnet server on the Windows machine and run a command line tool to check the current screen mode while StarCraft is running. I have used a couple of command-line tools for changing the screen mode [1] [2], but neither of them have an option to list the current mode. Does anyone know of a tool or can anyone quickly create a tool that will list the resolution currently in use from the command line?
I have tried using StraceNT [3] to see the parameters passed to ChangeDisplaySettingsEx (which is the function StarCraft uses to change the resolution), but it does not let me look at the DEVMODE structure that is passed in. Does anyone know of a trace tool that would let me do that?
Based on my testing so far, there appears to be a discrepancy in how Windows and Wine set the refresh rate if none is specified on a resolution change. Wine always picks the first mode with a matching resolution, while Windows does not. My tests in Windows have shown that even though StarCraft asks for 640x480 and the first mode with that resolution has a 60 Hz refresh rate, Windows chooses a higher refresh rate. Unfortunately, I don't know exactly which refresh rate, but I can verify by sight that 60 Hz is not being used (the screen doesn't flicker as it would at 60 Hz). This discrepancy is why I wish to run further tests and acquire more precise information about the refresh rate being used.
I am performing this testing according to a request in a previous thread [4]. If you have any questions about my testing, please let me know.
Denver