Fixes warning C4530: C++ exception handler used, but unwind semantics are not enabled
by disabling exceptions in the STL since ScummVM doesn't support them
This is another warning of the use of memset to clear a non-trivial
structure / class. This can be removed since the structure is cleared by
the constructor which will be called by the "new" so clearing this again
is redundant.
Apart from some basic cleanup this commit reverts a somewhat unfortunate design decision I made. The Kyra/Hof/Lol PC-98 sound drivers shouldn't inherit from the emulator. This commit separates the driver from the emulator putting some common interface in between. This should allow easier implementation of other PC-98 sound drivers.
No need to explicitly set it to 1.0. (There may have been in an
earlier version, to avoid any possible rounding error. But if so,
that reason is long gone.)
The high-resolution videos play back at much lower volume than the
original ones. This adds hard-coded values for how much to amplify
each cutscene. It's all done by ear, and it does introduce some
clipping, but I think it should be acceptable.
Of course, it could also be a problem with the audio decoder, so
this may be the wrong approach entirely.
It also hads further support for playing AIFF files. However, the
Blorb files for Lurking Horror & Sherlock on the if-archive website
don't seem to be valid AIFF files, so sound doesn't play using them
(This is basically an unnecessary emulation of a hardware limitation. EOB II will try to load approximately 70 KByte of samples for the outro sequence, which would lead to some missing sounds with the 64 Kbyte limit).
The function signature for these functions was changed from (char *) to
(const char *) in the v1.1.7 release, so compiling against
Fluidsynth v1.1.6 or earlier requires the copying of the strings to
prevent compilation errors such as "error: invalid conversion from
'const char*' to 'char*'".
Normally, we would break compatibility with older versions as platforms
should be using the latest Fluidsynth v1.X release of v1.1.11.
However, since this is trivial to fix and prevents breakage for legacy
platforms, am restoring the string duplication with scumm_strdup().
Apart from this, we should look at the Fluidsynth v2.X releases
currently in RC testing as the API is now changed for this.
As discussed on the mailing list we should use title capitalization
only for push buttons and tabs and use sentence capitalization for
everything else.
The OPL2LPT is an OPL2 chip plugged on a parallel port. It is
write-only but otherwise acts as a classic AdLib. This commit adds
support for this device.
User is expected to have the right permissions on the parallel
port. By default, the first suitable parallel port is used. It is
possible to override that with the hidden configuration setting
"opl2lpt_parport".
It depends on the presence of the libieee1284 library which abstracts
a bit parallel port handling. An alternative would be to access
directly /dev/parportX on Linux. This would amount of code but it
would be Linux-only.
Tested with Indy 3 and SOMI.
Previously, a C-style cast was used to convert a
Common::Array<Plugin *>, populated with pointers to StaticPlugin
and DynamicPlugin instances, to a
Common::Array<PluginSubclass<T> *>, but PluginSubclass<T> is a
*sibling* class to StaticPlugin/DynamicPlugin, so this cast was
invalid and the results undefined. The methods for retrieving
subclasses of plugins can't be easily changed to just generate an
array of temporary wrapper objects that expose an identical API
which dereferences to the preferred PluginObject subclass because
pointers to these objects are retained by other parts of ScummVM,
so the wrappers would needed to be persisted or they would need to
just re-expose the underlying Plugin object again. This indicated
that a way to solve this problem is to have the callers receive
Plugin objects and get the PluginObject from the Plugin by
explicitly stating their desired type, in a similar manner to
std::get(std::variant), so that the pattern used by this patch to
solve the problem.
Closes gh-1051.
If the audio thread called to readBuffer before any packet had been
added to the stream, the state of the stream would be changed from
INIT to EOS. Later, when a packet was received, the state would go
directly from EOS to READY, skipping decoder init, leaving garbage
memory in the decoder structs and causing a crash of the decoder.
Fixes Trac#9653.
