Currently, the text-to-speech manager will try to update the voices whenever TextToSpeechManager::pushState()/popState() is called. This causes lags of 4 - 5 seconds on Windows. Also, a warning is triggered each time. This commit prevents that from happening if the tts option is not enabled.
This commit currently affects only Windows. Other backends don't make use of the new _enabled setting. I don't know if it would make sense for any of these and I also wouldn't be able to test it.
This likely needs further work after some testing with OpenPandora toolchain
to check that this builds a working binary and a basic test to see if this
fixes the reported issue which has been present after v1.7.0 release which
did not exhibit this problem.
OGLES2 doesn't support uniform booleans so we use macros to make it use
integers in this case. For other platforms, this change should be a
noop.
OGLES2 doesn't like float suffix for constants, remove it as well.
We now link with OpenGL dynamically so symbols are provided by GLAD.
If it's not supported pointer will be null.
OGLES2 should support these functions so they must be removed
The previous workaround was introduced to fix fighting using the keypad
in Indy 3. It enforced genering a number on KEYCODE_KP# events even
when KBD_NUM is not set. The issue on AmigaOS is that SDL never reports
KBD_NUM as being set. Instead we get different keycodes depending if
numlock is on or off (e.g. KEYCODE_KP3 and KEYCODE_PAGEDOWN).
The new workaround is to set the KBD_NUM modifier whenever we receive
a KEYCODE_KP# event from SDL. This way we also generate a number, but
in addition this is consistent with the modifier and works with code
that checks the KDB_NUM modifier (such as the GUI code or the AGI
engine).
There are two ways the user can specify a screenshot path: by
editing the config file manually, or by passing it on the command
line. In the later case it is added to the transient domain that
is cleared when opening the launcher, so it only worked when also
specifying a game to start on the command line. With this change
a screenshot path specified on the command line will be used until
quitting ScummVM.
This could be confusing if the user had the ability to specify the
path in the ScummVM Options, as then we would probably want to use
the new specified path immediately. But since the path does not
appear in the options, this change should work fine.
Instead we call the OSystem_SDL implementation of getScreenshotsPath(),
as done in the POSIX backend. This change means that if we change how
we handle a user-specified screenshot path in the SDL backend, the
Windows and macOS backends will still get the correct path.
In HiDPI mode, the window coordinates and the drawable area coordinates
might be on a different scale. The allowed mouse area needs to be scaled
accordingly before passing it to SDL_SetWindowMouseRect().
This fixes bug #13152.
The mouse position is set in window coordinates, but it's clipped to the
game area in drawable area coordinates. Previously, the scaling between
these two was not taken into account when calculating the right/bottom
edges of the game area. When the clipped mouse position was converted
back to window coordinates and rounded to the nearest integer, it could
end up on the edge of the game area, not inside of it. This leads to a
loop in which the clipped mouse position is outside of the game area and
needs to be clipped again.
This fixes bug #12646.
Apple's desktop operating system was formerly called "Mac OS X" and "OS X", but since 2016 it has been called "macOS" (starting with version 10.12).
Changing across all comments and documentation to use this current terminology, except in cases where the historical versions are explicitly referenced. No code changes are made; we should consider changing those in future PRs.
