From: David B. <dav...@gm...> - 2011-04-11 17:39:21
|
Hi, > I think t4k_common-0.1.1 is ready for release, and tuxmath 2.0.0 (as I > finally decided to number it) is very close. The "new-art" images look > great, and the first set has been merged into master. The windows > crossbuild builds correctly and AFAICT works fine, including network > play. After some more testing with the windows build, it appears it really isn't ready for general consumption, although the Linux build seems ready. I'm seeing two main problems with the windows build: 1. In LAN play, there are too many dropped questions, i.e. when the server sends a question, but for some reason it doesn't appear on the client's screen. This leads to a sort of deadlock where the client is waiting for another question, and the server is waiting for an answer. This used to happen all the time when we were first getting the LAN game implemented. I thought I had eliminated this problem, and haven't seen it recently in Linux, but there must still be some sort of race condition allowing this to occur. 2. When using mouse control for the new Factoroids, the "game" ends unexpectedly after a few seconds, taking the player to the high score name entry. So something is causing the main game loop to exit, probably some error condition related to the mouse. Also, the mouse sensitivity seems insanely high in the windows build. So, I think I'll make a source-only 2.0.0 release, and hold off on a win32 binary until 2.0.1 (i.e. when we get the above problems fixed). Also, if I have a 2.0.0 source release posted, I can start working on updating tuxmath in MacPorts. Also, I'm trying to get t4k_common included upstream in mingw-cross-env, so that those who want prepare a crossbuild setup just have to install mingw-cross-env and type "make t4k_common". Once that is done, go to tuxmath's buildw32/ directory and run "./tmwin.sh". I want to make it really easy for members of our team to create win32 packages for testing. Best, David |