but at least it doesn't seem to do any harm.
Disabled the sound FX "garbage collection" in FxServer(). I'm not really
convinced it's necessary at all, and even if it is, doing it from a
separate thread it just begging for trouble. I've modified OpenFx()
slightly to deal with this, but I may still have introduced regressions.
Temporarily disabled the "goto label1" hack, since it seems to be the main
reason for ScummVM crashing if I allow a piece of music to finish on its
own (i.e. when not terminating it prematurely by triggering another piece
of music).
svn-id: r9990
the functions which access data manipulated by FxServer().
For instance, FxServer() may free bufferFx[i], which sounds potentially
unhealthy to me.
svn-id: r9989
channel indexes, we should use stopHandle() instead of stop() to kill the
music channel.
Am I the only one who finds the distinction between channel indexes and
sound handles confusing at times? :-)
svn-id: r9967
music channel has faded out, destroy the channel immediately. Don't wait
for the mixer to finish it off.
This seems to fix a problem where the mixer would eventually run out of
slots if you left the Quit dialog showing for too long.
Unfortunately I don't know if it fixes the "out of slots" errors I
encountered once during normal play. Oh well, time will tell...
svn-id: r9942
is the only place I can think of where this could have happened, so I've
added a paranoid check to ensure the buffer length is even.
Let's see how that works out...
svn-id: r9933
already reached its scroll target. This keeps BS2 from using all available
CPU time all of the time.
It may still be that we need a mechanism for throttling the frame rate when
the scene is moving towards a scroll target, but my computer isn't really
fast enough to test that.
Two other bugs fixed in the process:
* I think the last frame of the render cycle was rendered, but not
displayed. If so, that should be fixed now.
* I discovered that there are cases where we do need to clear the screen
(e.g. at the "Meanwhile..." message when George has found out about the
Glease Gallery), so I've re-enabled the function and disabled it in the
render cycle.
svn-id: r9904
block surfaces. (A block surface is a 64x64 tile of a parallax layer.)
I've also done a few things to try and optimize the drawing:
* The back buffer is no longer cleared between frames. This may cause
regressions, but I do believe that the entire picture area is always
completely re-rendered for each frame.
As a result of this, the menu code is now responsible for clearing the
icon areas itself.
* A few unnecessary copy_rect() calls were commented out in favor of one
big copy_rect() in ServiceWindows().
* Completely opaque block surfaces are copied with memcpy(), one line at a
time.
Unless we manage to add intelligent screen redrawing, I don't think it will
get that much faster than this, though there is some unnecessary data
copying in DrawSprite() that could be removed.
And the game is still a terrible CPU hog. I believe the animation runs at
approximately 12 fps. If there's still time left, it will pump out further
frames to get smooth scrolling. We ought to put a cap on that, and if it
has already reached the scroll target it should sleep for the rest of the
render cycle.
svn-id: r9886
I don't know if I got it right - the result doesn't look that great to me -
but at least the infrastructure is there.
This, I think, marks the point where BS2 graphics is pretty much done. Some
functions haven't been unstubbed yet, but I believe they're used for
debugging and/or profiling. I'm not sure they're worth the trouble.
Of course, there is still testing and clean-ups to make. For instance, I'd
like DrawSprite() to use malloc() a bit less.
svn-id: r9879
only uploaded to the backend if fadeNow == RDPAL_INSTANT. Otherwise, assume
that FadeServer() will do it for us, eventually.
I think this is the correct behaviour, and it prevents the bug where the
unfaded room image might flash by briefly when moving the mouse while
changing rooms.
svn-id: r9868
sprites are drawn, but I think that's how it should be.
1: No bells or whistles.
2: This setting adds sprite blending, e.g. the smoke at the docks or the
display cases at the Glease Gallery.
3: This setting adds light map support, e.g. when walking under the shack
at the docks.
4: This setting adds better scaling algorithms.
The first three settings should work fine now. In fact, the third setting
is what we used to implement. The fourth setting still needs work and
testing. I've added code for downscaling case, but frankly I'm not
convinced the result is any better than with the simpler scaler. I usually
can't even tell the difference.
Of course, my translation of the original code could very well be buggy.
svn-id: r9867
for the making it a timer handler. This should eliminate the occasional
glitches I've been seeing with fades not being completed.
I'm also hoping that it will fix the problem where the game would sometimes
hang when moving between rooms. I know that at least once when I had that
happen to me the game was busy-waiting for the palette to fade.
At the very least, it's one place less to worry about thread-safety in.
svn-id: r9854
unnecessary stuff from our own Surface class. The former allows the in-game
dialogs to at least sort of work, and the latter gained me a few frames per
second, according to the built-in FPS counter.
svn-id: r9825
under the shack), plus some other cleanups. The s->blend & 0x02 case looks
bogus to me, but I don't know where it's used and I can't see that the
original did it differently.
svn-id: r9819
the old way would have made it impossible to access the (not yet
implemented) menu icons at the top of the window. I hope this one works...
svn-id: r9793
comment, noting that DrawMouse() used to handle the "luggage animations" as
well. I don't know if I broke anything when I moved that part of the code.
svn-id: r9792
list of music types supported. This was done because now
PC speaker support must be treated separately, along with
Adlib and native (GM/MT32) support.
This fixes a problem with games that don't support PC speaker
(including V5 games that don't parse SPK resources yet)
being run with -epcspk or -epcjr. Those games now properly
switch to -enull so that music resources still get parsed
and music/script synchronization mechanisms don't break.
svn-id: r9703