README.QNX
author Sam Lantinga <slouken@libsdl.org>
Thu, 02 Apr 2009 04:43:36 +0000
branchSDL-1.2
changeset 4167 a6f635e5eaa6
parent 886 05c551e5bc64
child 3083 0bc41e0361d3
permissions -rw-r--r--
Fixed bug #611

From Tim Angus 2008-08-12 11:18:06

I'm one of the maintainers of ioquake3.org, an updated version of the
Quake 3 engine. Relatively recently, we moved ioq3 to use SDL as a
replacement for 95% of the platform specific code that was there. On the
whole it's doing a great job but unfortunately since the move we've been
getting complaints about the quality of the mouse input on the Windows
platform to the point where for many the game is unplayable. Put in
other terms, the current stable SDL 1.2 is basically not fit for purpose
if you need high quality mouse input as you do in a first person shooter.

Over the weekend I decided to pull my finger out and actually figure out
what's going on. There are basically two major problems. Firstly, when
using the "windib" driver, mouse input is gathered via the WM_MOUSEMOVE
message. Googling for this indicates that often this is known to result
in "spurious" and/or "missing" mouse movement events; this is the
primary cause of the poor mouse input. The second problem is that the
"directx" driver does not work at all in combination with OpenGL meaning
that you can't use DirectInput if your application also uses OpenGL. In
other words you're locked into using the "windib" driver and its poor
mouse input.

In order to address these problems I've done the following:

* Remove WM_MOUSEMOVE based motion event generation and replace with
calls to GetCursorPos which seems much more reliable. In order to
achieve this I've moved mouse motion out into a separate function that
is called once per DIB_PumpEvents.

* Remove the restriction on the "directx" driver being inoperable in
combination with OpenGL. There is a bug for this issues that I've
hijacked to a certain extent
(http://bugzilla.libsdl.org/show_bug.cgi?id=265). I'm the first to admit
I don't really understand why this restriction is there in the first
place. The commit message for the bug fix that introduced this
restriction (r581) isn't very elaborate and I couldn't see any other bug
tracking the issue. If anyone has more information on the bug that was
avoided by r581 it would be helpful as I/someone could then look into
addressing the problem without disabling the "directx" driver.

* I've also removed the restriction on not being allowed to use
DirectInput in windowed mode. I couldn't see any reason for this, at
least not from our perspective. I have my suspicions that it'll be
something like matching up the cursor with the mouse coordinates...

* I bumped up the DirectInput API used to version 7 in order to get
access to mouse buttons 4-7. I've had to inject a little bit of the DX7
headers into SDL there as the MinGW ones aren't up to date in this respect.
     1 README.QNX by Mike Gorchak <mike@malva.ua>, <lestat@i.com.ua>
     2 Last changed at 24 Apr 2004.
     3 
     4 ======================================================================
     5 Table of Contents:
     6 
     7 1. OpenGL.
     8 2. Wheel and multi-button mouses.
     9 3. CDROM handling issues.
    10 4. Hardware video overlays.
    11 5. Shared library building.
    12 6. Some building issues.
    13 7. Environment variables.
    14 
    15 ======================================================================
    16 1. OpenGL:
    17 
    18     OpenGL works well and is stable, but fullscreen mode has not  been
    19 heavily tested yet.
    20     If you have QNX RtP version 6.1.0 or above you must  download  the
    21 Photon3D runtime from http://developers.qnx.com or install it from the
    22 public repository or from the public CD, available with QNX. OS versi-
    23 ons below 6.1.0 are not supported.
    24     When creating an OpenGL context, software renderer mode is artifi-
    25 cially selected (QSSL made acceleration  only  for  Voodoo  boards  in
    26 fullscreen mode, sorry but I don't have this board  to  test  OpenGL -
    27 maybe it works or maybe not :)). If you  want  acceleration - you  can
    28 remove one line in the source code: find the  file SDL_ph_image.c  and
    29 remove the following
    30 
    31     OGLAttrib[OGLargc++]=PHOGL_ATTRIB_FORCE_SW;
    32      
    33 line in the ph_SetupOpenGLContext() function or change the argument to
    34 PHOGL_ATTRIB_FORCE_HW or PHOGL_ATTRIB_FAVOR_HW.
    35 
    36 ======================================================================
    37 2. Wheel and multi-button mouses:
    38 
    39     Photon emits  keyboard  events (key up and  down)  when the  mouse
    40 wheel is moved. The key_scan field appears valid, and it contains zero.
    41 That is a basic method of detecting  mouse  wheel events under Photon.
    42 It looks like a hack, but it  works for me :) on various PC configura-
    43 tions.
    44 
    45 I've tested it on:
    46 
    47 1. Genius Optical NetScroll/+ PS/2 (1 wheel)
    48 2. A4Tech Optical GreatEye WheelMouse PS/2,  model: WOP-35.  (2 wheels
    49    + 2 additional buttons). The wheel for vertical scrolling  works as
    50    usual, but the second wheel for horizontal scrolling emits  two se-
    51    quential events up or down, so it can provide faster scrolling than
    52    the first wheel. Additional buttons don't emit  any  events, but it
    53    looks like they're handled by photon in an unusual way - like click
    54    to front, but works not with any window, looks like a fun bug-o-fe-
    55    ature :).
    56 
    57 ======================================================================
    58 3. CDROM handling issues:
    59 
    60     Access to CDROM can only be  provided  with  'root'  privileges. I
    61 can't do anything about that, /dev/cd0 has brw------- permissions  and
    62 root:root rights.
    63 
    64 ======================================================================
    65 4. Hardware video overlays:
    66 
    67     Overlays can flicker  during  window  movement,  resizing, etc. It
    68 happens because the photon driver updates the real window contents be-
    69 hind the overlay, then draws the temporary chroma key color  over  the
    70 window contents. It can be done without using the chroma key but  that
    71 causes the overlay to always be on top. So  flickering  during  window
    72 movement is preferred instead.
    73     Double buffering code is temporarily disabled in the photon driver
    74 code, because on my GF2-MX it can accidentally cause a buffer  switch,
    75 which causes the old frame to show. S3 Savage4 has the  same  problem,
    76 but ATI Rage 128 doesn't. I think it can be fixed later. Current  code
    77 works very well, so maybe double buffering is not needed right now.
    78     Something strange happens when you try to move the window with the
    79 overlay beyond the left border  of the screen. The  overlay  tries  to
    80 stay at position x=0, but when attempting to move it  a  bit  more  it
    81 jumps to position x=-60 (on GF2-MX, on ATI Rage128 this  value  a  bit
    82 smaller). It's really strange, looks  like  the  overlay  doesn't like
    83 negative coordinates.
    84 
    85 =======================================================================
    86 5. Shared library building:
    87 
    88     A shared library can be built, but before running  the  autogen.sh
    89 script you must manually delete the libtool.m4 stuff from the acinclu-
    90 de.m4 file (it comes after the ESD detection code up to the end of the
    91 file), because the libtool stuff in the acinclude.m4 file was very old
    92 in SDL distribution before the version 1.2.7 and doesn't knew anything
    93 about QNX. SDL 1.2.7 distribution contains the  new libtool.m4 script,
    94 but anyway  it  is  broken :), Just  remove  it, then  run "libtoolize
    95 --force --copy",  delete the file aclocal.m4 if it is exists and after
    96 that run the  autogen.sh script. SDL 1.2.8 contains  fixed libtool.m4,
    97 ltmain.sh and config.sub files, so you  can  just  run  the autogen.sh
    98 script.
    99 
   100 ======================================================================
   101 6. Some building issues:
   102 
   103     Feel free to not use the --disable-shared configure option if you'
   104 ve read the above comment about 'Shared  library  building'. Otherwise
   105 this  option  is  strongly  recommended, as  without it the sdl-config
   106 script will be broken.
   107 
   108     Run the configure script without x11 support, e.g.:
   109 
   110     a) for OpenGL support:
   111     ./configure --prefix=/usr           \
   112                 --disable-video-x11     \
   113                 --disable-shared
   114 
   115     b) without OpenGL support:
   116     ./configure --prefix=/usr           \
   117                 --disable-video-x11     \
   118                 --disable-shared        \
   119                 --disable-video-opengl
   120 
   121     And of course dont forget to specify --disable-debug, which  is on
   122 by default, to disable debug and enable the expensive optimizations.
   123 
   124     In the test directory also run the ./configure script without
   125 x11 support, e.g.:
   126 
   127     ./configure  --with-sdl-prefix=/usr            \
   128                  --with-sdl-exec-prefix=/usr       \
   129                  --prefix=/usr --without-x
   130 
   131 ======================================================================
   132 7. Environment variables:
   133 
   134     Please note that the photon driver  is  sensible to the  following
   135 environmental variables:
   136 
   137  * SDL_PHOTON_FULLSCREEN_REFRESH - this environment variable  controls
   138 the refresh  rate  in  all  fullscreen  modes. Be  carefull !!! Photon
   139 drivers usually do not checking the maximum  refresh rate, which video
   140 adapter or monitor supports.
   141 
   142  * SDL_VIDEO_WINDOW_POS - can be  set in the "X,Y" format.  If X and Y
   143 coordinates are bigger than the current desktop resolution, then  win-
   144 dow positioning across  virtual  consoles is activated. If X and Y are
   145 smaller than the desktop resolution  then  window  positioning  in the
   146 current console is activated. The word "center" can be used instead of
   147 coordinates, it  produces  the  same  behavior  as  SDL_VIDEO_CENTERED
   148 environmental variable.
   149 
   150  * SDL_VIDEO_CENTERED - if this environmental variable exists then the
   151 window centering is perfomed in the current virtual console.
   152 
   153 Notes: The SDL_VIDEO_CENTERED enviromental  variable  has greater pri-
   154 ority than the SDL_VIDEO_WINDOW_POS in case if both variables are sup-
   155 plied to the application.