recently added (cursor) palette manager) so that the cursor can be properly
restored after returning from the GUI.
If there's any C++ magic that can keep the backend functions from being called
by anything else than these managing classes, that would probably be a good
idea. Also, since the cursor manager keeps a copy of the cursor image, perhaps
there are at least some backends that will no longer need to?
svn-id: r22639
- Eriktorbjorn's patch from same tracker item for scaling sword1/2 cutscenes
is applied as is. It lacks resolution switch on-the-fly.
- GUI widgets are repositioned on the fly and use most space, even aspect
ratio corrected screen is used without scaling
- Heavy tesing is required, but works for me in all cases except for bug
#1483272: "GUI: SCUMM pause dialog breaks upon scaler switch" which needs more
work.
- I probavly broke some backend or two
svn-id: r22505
* Doing this, split scummInit into setupScumm and resetScumm methods
* Moved some init code to new methods (huuuuge init methods are hard to read, and hard to customize using subclassing)
* Renmaed various setup???Palette methods to set???Palette
* Some minor cleanup / tweaks
svn-id: r22227
- Added Engine::_targetName whose value is computed from the name of the active domain
- Removed GameDetector::_targetName, instead code now uses either Engine::_targetName or the name of the active domain
- This in turn allowed for removing usage of GameDetector in many places
svn-id: r21916
- This also means you can now set those options in the config file
- Fixed a bug I recently introduced that made bool command line options (like -u, -f) always return 'false' when used in their single letter form
svn-id: r21909
Fix a lot of debug/error statements that were using %d/%x
for the result of pointer arithmetic.
As C++ apparently has no format string for ptrdiff_t use
%lu/%lx as appropriate.
svn-id: r21824
* IMuse is no longer a concurrency front-end.
* IMuseInternal now derives from IMuse.
* Common::StackLock used to protect thread-sensitive interface methods (same as IMuseDigital).
* clear_queue() included in stopAllSounds() so it can be removed from the public interface.
* Game ID now specified at init using property().
* Timer callbacks receive a struct containing IMuseInternal and MidiDriver refs, instead of just the latter.
* OSystem pointer from init is now cached and used instead of global.
All references to the g_system and g_scumm globals are now gone. BOOYAH!
Tested with MI2, DOTT and S&M, under Windows, in Native MIDI, Adlib, and Mixed modes. No regressions or concurrency issues observed. Manifestations of the latter are the biggest concern at this point.
svn-id: r21104
