This also should fix some regressions from the previous
commits, related to MidiParser's either being leaked,
or being deleted and then used again (i.e., crashing).
I tested as many games as I had available, but further
testing of all affected engines is called for anyway.
The regression affected AGOS and maybe some others; specifically,
the real MidiDriver would have been deleted twice -- I previously
missed that the Engine instances takes care of freeing the real
MidiDriver, not the MidiPlayer wrapping it.
This commit should clarify the ownership of the real MidiDriver for
most pseudo MidiDrivers.
Many engines follow the advice in audio/midiparser.h and create a
"pseudo-MidiDriver" subclass. But MidiParser really only needs a tiny
subset of the MidiDriver capabilities, namely those found in
MidiDriver_BASE. So we try to subclass from that whenever possible; this
allows us to remove many stub methods, and enables further future
simplifications.
This in turn enables modifying MidiDriver_MPU401::close() to allow
it to be called on a midi driver that has not yet been opened.
The specific issue that triggered me to make these changes was a
crash-upon-quit in HUGO, caused by it instantiating a midi driver,
then encountering an error (missing hugo.dat) *before* having
opened the new midi driver; the general cleanup code then tries
to close the (not yet opened) midi driver -> kaboom
Also fixed some engines which were leaking MidiDriver instances.
This is done by converting the internal RGBA palette data to RGB before
calling setPalette. This might not be the best solution, but looking a bit
into the engine it seems like changing all the code to work with RGB instead
of RGBA might require some bit of work.
This is a regression from r55256. Apparently, SCUMM has issues when sending a sustain
off on a notes off event. Thus, this has been turned into a feature, which is disabled
by default. Since MADE, SAGA and tinsel all share the same music code and play regular
MIDI files, and this feature fixes hanging notes for them, it has been enabled for them.
Also, applied a patch for a bug regarding the notes off event in MADE and tinsel,
which has been applied in SAGA already
svn-id: r55746
This fixes the leak from the COR0_BEGIN_CODE macro's CoroContextTag allocation.
Setting the _sleep value to negative values prevented the deallocation of these.
Have tested for obvious regressions, but a full DW/DW2 playtest should be done to confirm that this doesn't cause any issues.
Thanks to fingolfin for indicating this solution.
svn-id: r55671
When copying data files with 8.3 names from a CD, they sometimes end up
with a trailing dot, e.g. "INDEX." instead of "INDEX". Comon::File
supports this, but in r54392 Tinsel stopped using Common::File.
This commit also removes the useless SeekableSubReadStream wrapper
around the actual file stream.
svn-id: r54446
Use of global vars is what prevents RTL from working in Tinsel (and
probably in other engines). More specifically, the fact that many
global vars are not explicitly inited when the engine is (re)launched.
svn-id: r54262
This commit contains the AudioCDManager changes from the gsoc2010-opengl
branch. The other changes in that branch are restricted to the backends
directory only (plus configure).
The Nintendo DS and Dreamcast ports still need to be ported over to
the new Audio CD system, but that should be fairly easy to do.
svn-id: r54147
* names now comply to our naming conventions
* the function computeStreamMD5AsString which computes the MD5
as a hex string now returns it as a Common::String
* add doxygen comments
svn-id: r54121
* Fix a bug which caused coroutines invoked with nullContext to
leak their state.
Sadly, nullContext is underdocumented, yet very delicate and
full of subtleties... ./
* Move nullContext decl from sched.cpp to coroutine.cpp
* Enhance a few doxygen comments
svn-id: r53934
This fixes some leaks that occurred when exiting the game resp.
returning to launcher. Note that we still leak some coroutines when
exiting after loading a savegame.
svn-id: r53933
