MKID_BE relied on unspecified behavior of the C++ compiler,
and as such was always a bit unsafe. The new MKTAG macro
is slightly less elegant, but does no longer depend on the
behavior of the compiler.
Inspired by FFmpeg, which has an almost identical macro.
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.
Some dialogue answers are set to the literal 'null', which is used in NS to
simulate a fixed monologue ('null' instructs the engine to jump to the first
entry in the list of available questions, which is always made of a single
item in this case).
BRA has flexible monologues that depend on the game state, so this patch goes
and checks all the available branches before picking the correct one.
svn-id: r55591
This will now avoid invalid memory reads and instead emit warnings if it is called with values outside of the expected data buffer or on a NULL buffer.
svn-id: r55492
* Renamed and retyped "int _fieldC" to "bool _stillWalkingTowardsNode" to aid readability and clarity of code. Thanks to fuzzie for naming.
* Added assertion to prevent use of invalid memory as reported by Valgrind when walk point p is constructed from an empty list.
* Added code to stop walk if buildPath fails, though this causes an assertion instead currently.
svn-id: r55279
This command forces the mouse state to Enabled and Visible.
This is intended to help playtesting in Big Red Adventure, which has an issue currently with the mouse pointer getting stuck in the hidden state in some cases.
svn-id: r55234
The automatic destruction of engine member "_char" of type Character at end of the destructor causes an invalid read in the destruction of the AnimationPtr which is a SharedPtr. To stop this happening, the SharedPtr is reset() prior to the destruction of Gfx, so it's dependencies on Gfx are still present. This will affect Nippon Safes as well as Big Red Adventure, but is unlikely to cause any issues and it is likely this exists in NS as well.
Thanks to fuzzie for this code.
svn-id: r55233
This corrects the Valgrind invalid reads associated with the IS_PATH_CLEAR() check reading beyond the data buffer and probably improves the pathfinding behaviour.
Thanks to fuzzie for suggesting this patch.
svn-id: r55231
