• Advertisement


  • Content count

  • Joined

  • Last visited

Community Reputation

5443 Excellent


About Endurion

  • Rank

Personal Information

  1. Thanks for the help! I found my issues: a) There's a flag in ResizeBuffers, which overrides the previously set flags used in CreateSwapChain. I added DXGI_SWAP_CHAIN_FLAG_ALLOW_MODE_SWITCH there as well (as I did in CreateSwapChain). In conclusion: Make sure both flag values match! b) I'm listening to WM_ACTIVATEAPP now, and when I'm losing focus while in fullscreen I'll switch to windowed mode.
  2. Multiplayer Pac-man Challenge

    Added a project page here: Can't wait to see other entries!
  3. I'm running Windows 10, latest official build. I've tried not to touch WndProc in the DX11 renderer, but I'll probably do that as a last resort.
  4. I've recently found ugly problems with the d3d11 renderer in my game engine. I'm using IDXGISwapChain with real fullscreen and I have serious problems getting Alt-Tabbing to work. Switching between windowed and full screen mode works completely fine. That was easy to do with DX8/DX9, where didn't have to do anything but check for a lost device. Alt-tabbing from fullscreen restored the display mode and everything worked fine. Now with the IDXGISwapChain I can't alt-tab out of a game running in fullscreen properly anymore. Upon Alt-tabbing the display goes bonkers (e.g. a part of the screen is displayed half offscreen, but also another part in the top left corner (looks like the resolution has changed, and some old graphics are still there in GPU memory). I do not see the desktop at all. Upon clicking anywhere the game is switched to windowed mode, but with the window stripped of all borders (and therefore the client size being not the size I'm expecting). I do not modify the window styles myself (as I did in DX8/DX9), so I suspect IDXGISwapChain. Unfortunately the extra debug output of D3D11 doesn't show any errors. I've called MakeWindowAssociation with the flag to not affecting anything (not listenen to Alt-Enter or switch modes). Does anybody have any ideas what I might be doing wrong? Or maybe have a working sample that does windowed and fullscreen toggling with Alt-tabbing working? Would reacting on WM_ACTIVATEAPP to switch to windowed mode myself be a good solution?
  5. Multiplayer Pac-man Challenge

    Just to be sure to have an entry up in time, here's my game, Get 'em Good! It's a clone of my C64 game, with almost completely original graphics and music, the latter not done by me. Play alone or in coop mode (one score and one extra pickup counter for both). Eat all dots, fight monsters, get the key and unlock the exit. Monster grow stronger and faster when advancing stages. Your only chance to survive is to kill enemies and gather pickups to get stronger yourself. To activate an extra gather 9 pieces of the same item. The player that collects the 9th extra gets the power up Most power ups stack, so keep collecting! The stages are randomly created, so every game is different. The download is here: Download Here Have fun and let me know if you experience problems! Edit: updated the archive with the game source (not the common lib code though) PS: Refrain from alt-tabbing in fullscreen. It's hilariously broken. If you reach that state you may want to delete the xtreme.cfg file in %APPDATA%/Roaming/GR Games/GetEmGood. Sorry! Aftermath: That full screen problem is really annoying. Everything that worked fine and was easy to do in DX8 and 9 is now obsolete and the behaviour of IDXGISwapChain is utterly broken. Whatever you do, never Alt-tab in full screen. It stuffs up the window.
  6. Multiplayer Pac-man Challenge

    I'm nearly done on the game. Just my DX11 renderer is acting up when alt-tabbing from Fullscreen. This IDXGISwapChain thing is butt ugly finicky.
  7. Multiplayer Pac-man Challenge

    Better than thought actually. I've decided to scrap the pure Pacman approach and basically port an existing game from C64 to PC (Get 'em DX)
  8. Multiplayer Pac-man Challenge

    I'm still stumbling. I've so concentrated on getting the original behaviour done I'm basically stuck Let's see if I can manage to get something playable...
  9. Multiplayer Pac-man Challenge

    Regarding the 2 player part, all three options or does one suffice?
  10. Multiplayer Pac-man Challenge

    Oooh, neat I'm pretty much in!
  11. No new challenge yet? Need proposals? Staying with the arcade themes: * Pacman * Tetris * Dig Dug
  12. As a hobby developer I'm also not particularly fond of the signing stuff, as I provide my games for free. So I don't see to shell any money out for anything, esp. not periodically. IMHO this is nothing about trust. Have you personally ever checked the signature of a executable you downloaded? The suspicious warning works in a way that the number of downloads are weighed against the number of user warnings via feedback, so over time the warning should go away (IIRC this is the "Smart Screen" filter) One method could be to release your game in the Windows store. It's a bit annoying to modify, but you can use C++ just fine. To use the store costs a one time fee (ca. 17€) which I'm fine with. Still, I provide most games from a zip archive, only a select few with an installer (those who cater to rather non technical users)
  13. I think scaling up with whole numbers and blacking the rest is the proper way to go. Then you'd have good looking graphics and use as much as possible of the monitors area. I prefer having black bars to skewed graphics.
  14. No. Not color picking. You do not read back from the GPU if you can avoid it. Memory access is extremely optimized in one direction, from CPU to GPU. Since you prepare the map texture yourself somewhere, have a simple color coded version of it in memory. Check the pixel the mouse is over, color = state.
  15. A Windows Store Story - Part #5

    Progress! I managed to get remote debugging working on the tablet and found the cause of the crash bug at once: Interestingly during initialisation of the input wrapper I registered for gyro changing events before the basic input setup was complete. And, lo and behold, I received an event before the input init completed. Null pointer, bam! instant crash. Weirdly, this never happened on Windows phone. Seems UWP has some subtle differences between platforms. So now that crash cause is history and the next release is up. Now a bit more about the framework I've setup. I've tried two or three times to set up a simple UWP sample app and include the main game code in that. The sample app was setup differently (but I think I didn't do it on purpose, rather the SP of Visual Studio 2015 changed to 3 in between) With the first two tries the sample app had some XAML stuff (which I don't really like), and I never got the swap chain to work properly. With the 3d try there was no XAML document at all, and this time the swap chain worked (after a few attempts with different parameters) There's quite a few preconditions and certain parameter combinations that are mandatory. Luckily the runtime spits out a pretty descriptive error message what was actually expected. Using a generic framework for both Win32 and UWP proves a tad tricky, since UWP wants a managed main method. With a bit of pseudo global magic (I've got one environment instance where several services like renderer, logger, sound and also the main application register) I was able to refactor the main function out. So for HitBlock Deluxe I have a solution with a shared project containing the actual game logic, and a UWP and Win32 application project; both having the shared project as dependency. Current gotchas: Since UWP requires assets to be embedded, and absolutely does not want assets from outside the project folder, I share the assets from the UWP project. Ugly, but works. And there's a new crash, unfortunately without any crash info. Oh well....
  • Advertisement