Originally authored by @magicmyth in PR #1039
ScummVM is optimised to render frames when something changes on screen.
As some host environments can perform poorly if the app does not
reliably refresh its output regularly (notably Steam overlay) this new
option ensures that Scummvm outputs a minimal amount of frames even if
nothing is changing in the game renderer.
Currently this is only implemented in the SDL OpenGL renderer.
The new config option is called force-frame-update and it takes a
integer value representing the desired minimum milliseconds Scummvm
should wait before forcing a screen update/refresh. E.g: 50.
Note that the rendering system will not force a re-draw of a frame if
the app has rendered a changed frame within the desired minimum refresh.
Thus if the app is outputting 30fps and force-frame-update is set to
100ms (~10fps) then no duplicate frame will be shown (in theory).
As this is implemented in OpenGLSdlGraphicsManager::updateScreen()
OpenGLGraphicsManager::_forceRedraw has had its access changed to
*protected* so that it can be access by it's descendant. The reason this
has been done in OpenGLSdlGraphicsManager::updateScreen() is so
SDL_GetTicks() can be used to track the elapsed time. If it is useful
for other platforms using OpenGL to have this feature it could be
implemented within OpenGLGraphicsManager::updateScreen() provided a
suitable platform independent replacement for SDL_GetTicks() is used.
This would potentially be better as OpenGLSdlGraphicsManager checks
various other states when deciding if the screen should update.
- Use a SVG asset to allow for better scalability
- Make the controller visible when touching the screen
- Allow for oblique moving by placing finger between two directions
- Add more buttons on the center area of the screen (GUIDE, START, LEFT
STICK, RIGHT STICK) and simplify right area (only four buttons)
- Don't track all fingers, only the ones needed
- Cleanup now unused code
To apply a mask to a RGB (non-paletted) cursor in SurfaceSdl, we use an
intermediary masked_image, where the pixels made transparent by the mask
are replaced by an ARGB color with alpha = 0 (ARGB 0x00000000 in practice).
However the cursor surface still has a key color attached to it by
SDL_SetColorKey (set to 0 currently), which is not sensitive to the alpha
channel, wrongly making opaque black ARGB 0xFF000000 transparent as well.
So we provide an option to disable the key color in such a case.
- fixed aspect ratio correction (overlay destroyed
its state, keyboard shortcut didn't always work),
as well as significantly sped it up on VGA and
SuperVidel
- smoother transitions between video modes
- reduce the number of video mode changes
- fixed a few extreme cases when triple buffering
could lose an update
- lighter ST RAM usage in the lite build
- removed hardware accessing init/deinit routines
Instead of having one globally in AtariGraphicsManager. This simplifies
code and cleans up a lot of dark corners in AtariGraphicsManager.
A slight disadvantage is that now the cursor surface has be to converted
for each buffer.
This unifies the ImGui processing for both 2D and 3D engine.
This also make sure we don't destroy ImGui if init failed (like on older
OpenGL).
The destroy callback is also called whenever we destroy the context.
Only do it when ImGui is enabled by the engine and move the check in SDL
backend where ImGui is supported.
This makes the OpenGL backend independent of ImGui.
It's not always set correctly by engines (AGS) and a fill call to clear
the screen will set it to 0 along with the other components.
Setting the framebuffer to some transparent value is not something we
intend to do as this let the content underneath the drawing texture
visible like it's the case on ChromeOS.
This has been reported here: https://forums.scummvm.org/viewtopic.php?p=99293
- every malloc is now aligned on 16 bytes
- no need for the hidden pointers
- ct60_vmalloc() is called only once, the rest is handled
by dlmalloc (SV XBIOS' vmalloc is pretty fragile)
Otherwise, the cursor scaler and standard scaler may be different types,
which can lead to scaling the cursor with a scaler that normally doesn't
allow it. In this particular bug report, it caused ScummVM to crash.
_offsettedSurf must have the original screen dimensions, similar as done in
AtariGraphicsManager::Screen::reset().
Strangely, this bug was reproducible only in MagiC and not in FreeMiNT+MP.
Le Philousophe
Eugene Sandulenko
Vladimir Serbinenko
Sebastien Ronsse
Hubert Maier
Filippos Karapetis
Donovan Watteau
Simon Delamarre
Thorsten Otto
Miro Kropacek
Cameron Cawley
scemino
Matthew Jimenez
Christian Kündig
Torbjörn Andersson