This class manages a buffer where all data usually stored in detection
plugin will get copied before unloading the plugin and starting the
game.
This class expects that two functions are present in every
GameDescription: sizeBuffer which calculates how many bytes we will need
to store the entry in RAM and toBuffer which copies the data in the
buffer and fix the pointers in the class.
At the end, it is expected that an ADDynamicGameDescription doesn't
depend anymore on data stored in the detection plugin.
The AD_GAME_DESCRIPTION_HELPERS macro allow to implement these functions
in all GameDescription which don't have any pointer except those in
ADGameDescription.
I don't have the floppy version of Simon 2, and I'm uncertain of the
actual effect of failing the copy protection in Waxworks and Simon 1.
But they are at the very least partly implemented.
It can be reproduced by clicking on the crystal ball icon.
From looking at disasm I think it happens in the original
in the same way. But this is a static memory area, so it
will always corrupt the same variable in the original
(_fcsData1[0]) which seems to cause no harm here.
("Waxworks crashing at Egypt Level 3, corrupting save file")
The design for the timers is a bit sloppy, even the original
interpreter often writes negative timeouts into the save files.
We replicated this correctly, but treated these values as very
large positives which effectively disabled the timers.
The bug report is only about Waxworks, but I checked Elvira
and i have the same absurd timer values in my old savegames
there. It seems that the timers have little impact in most cases,
but in Waxworks there is a timer that runs every 10 seconds
which cleans up the items chain. And if that doesn't happen
it will lead to invalid item handles, like in that bug ticket.
This one is pretty obvious when comparing ScummVM
with DOSBox behaviour. After the cursor changes to
one of the arrow cursors, it often stays that way when
it should actually switch back to the normal cursor. I have
fixed it to match the original code.
zlib is used when available and falls back on gzio.
This allows performance improvements as our CRC32 and gzio
implementations are slower than base zlib.
As zlib is available when libpng is present, this is sensible to
benefit from it.
This change introduced several character animation regressions in the
game, which may seem as a worse experience than the original issue with
the pirate dialogue, for most end users (Trac#13997).
This reverts commit eb851041ed.
This adds detection entries for the German and French versions. All language
versions except Russian use WAV sound effects and GM SMF MIDI data, so the
platform is changed to Windows. These versions also contain the MT-32 MIDI data
in XMIDI format, so support for this is added as well. Finally, the MT-32 track
10 workaround has been updated to deal with these versions.
Adds detection entries for the new versions in the 25th Anniversary editions.
The remaining language versions are already detected.
To use the 25th Anniversary versions you have to copy the files from the
subdirectory of the language you want to use to the main directory, then detect
the main directory; or just detect without copying for English CD.
GOG has not included the AdLib instruments file with their release of Simon the
Sorcerer 1 DOS version. The game would not start in ScummVM when this file was
missing and AdLib was selected. Added a fallback to the generic AdLib driver
with built-in instruments.
GOG has not included the AdLib instruments file with their release of Simon the
Sorcerer 2 DOS version. The game would not start in ScummVM when this file was
missing and AdLib was selected. Added a fallback to the generic AdLib driver
with built-in instruments.
Also made the Windows version start with AdLib if an OPL2-only emulator is used.
When the noop flag was set on a MIDI event by the AGOS Simon 1 Windows or GMF
parsers, it would not be cleared when parsing the next event, leading to all
subsequent MIDI events being ignored.
Fixed this by setting the noop flag to false when it is not applicable to a
MIDI event.
When using this option with Simon 1, Simon wouldn't be cleared from the
bottom of the screen, if pressing Esc when Simon starts walking through
the screen right after the "Simon the Sorcerer" title is printed, at the
very beginning of the game.
Triggering the clearSurfaces() code even when _neverFade is enabled
solves this problem.
Fixes a mistake of mine in commit 802caa4c39.
