The functions loader_C and loader_Croot provably did not do anything
besides returning an error, so removing them was safe.
For loader_Lua, this was using luaL_loadfile. But that in turn now
uses Sword25FileProxy, which only supports loading config.lua,
and config.lua is not used via the loadlib mechanism.
Therefore, I deemed it safe to also remove this third loader.
This leaves loader_preload as only remaining loader. It is probably
unused, too, but I did not both to investigate further, as I already
achieved my primary goal (getting rid of fopen/fclose use).
All other removed functions were unused due to the removal of the
first three loader_* functions.
This just uses the provided filename rather than trying to recreate
it with a hard-coded target (causing crashes with other targets).
(Also, add an error check rather then crashing there, just in case.)
This is needed now, since changing the language in-game updates the ScummVM language setting, and there must be a matching detection entry for each language.
The trydecpoint() function *is* used, though probably only in
countries which don't use a decimal point. We can't use the ISO C
locale functions here because they're not fully implemented on
some platforms, e.g. Android. Hopefully this method will work.
This unifies the engine names in MetaEngine::getName() and the
credits. In particular drop "Engine" or "engine" from the names when
it was present and use expanded names in credits when the
MetaEngine uses it (e.g. "Beneath a Steel Sky" instead of "BASS").
- Savegame thumbnails are now created using a flat image format,
instead of PNG. This allows us to remove the PNG encoding code used
in the engine. Note that the saved games we create cannot be used
with the original engine anyway, so this change does not break
savegame compatibility with the original game (it's already broken).
- Compatibility is preserved with older saved games that were made with
ScummVM.
- Moved the embedded thumbnail reading code out of the PNG loader class,
to stop it from needlessly checking every PNG file.
- With this change, libpng is no longer required for the sword25 engine
I am not 100% sure whether the surfaces all use the same format as the screen,
but a quick test showed that it still works fine. In case this is wrong please
set them up with the correct format.
This should not cause any code behavior changes at this time, but if any
of the intermediate VideoDecoder classes ever starts to overload stuff,
this would become important.
svn-id: r55841
- Increase the resource cache limits
- Added a check before forcing resources to be freed
- Only force free image and animation resources, with a warning. It seems like there is
a bug in the resource reference code and several bitmap resources are not freed - added
a FIXME
- Clarify that initializeAnimationResource() is used with XML resources
svn-id: r55736
libpng is still needed for PNG encoding (for thumbnails in saved games of sword25), but
since we'll probably drop support for the original saved games anyway, the PNG encoding
code will ultimately be removed
svn-id: r55723
- Optimized and cleaned up translateYUVtoRGBA()
- Disabled a lot of seeking-related functionality
- Replaced some calloc() calls, used for audio buffer creation with malloc()
svn-id: r55666
Translated some comments, and pushed the indirect rendering define to the header
file, so that the engine won't try and update the screen with direct movie rendering.
Also, the thumbnail hack has been disabled, as it doesn't really work (at least not for
me: all the thumbnails are gray)
svn-id: r55663
The original checked the total amount of memory occupied by all resources. This has been
changed to a maximum number of simultaneous resources instead, so the game resources
are no longer leaked. Also disabled the unused or debug functions getUsedMemory(),
setMaxMemoryUsage(), setMaxMemoryUsage(), isLogCacheMiss(), setLogCacheMiss(). Performed
some cleanup on code related to the above.
svn-id: r55594
This reduced the initial memory used by 100MB for me, though the game keeps allocating
new resources in each scene without deleting them, because of the missing functionality in
getUsedMemory(). This change also slightly reduces the loading time on game startup.
svn-id: r55593
In the other cases I've found in the code, the colour components are
stored in the order B, G, R and A. Assume that's the case here too. I
hope that is correct. It doesn't seem to break anything obvious.
svn-id: r55255
This makes it possible to write
DECLARE_SINGLETON(foo);
instead of
DECLARE_SINGLETON(foo)
without causing a warning about an extra semicolon.
The extra semicolon helps some editors at parsing the C++ code.
svn-id: r54258
Since SWORD25 uses Debug Channels, this allows for the interactive setting of debugflags as well as providing a base for adding further debugging commands.
However, this is not currently usable as the command key code is commented out.
This is due to the event loop which reads keyboard input being buried 2 object layers below the Engine VM object and I am unsure how the engine development team would want this exposed / interfaced.
svn-id: r54184
A new header file common/forbidden.h is included by scummsys.h and it
re-#defines numerous symbols like fopen(), fread(), system(), etc. with
garbage, in order to provoke compile errors in any code using them.
If a .cpp file really *must* use any of these (e.g. because it is a
backend file), then these redefinitions can be disabled by #defining
FORBIDDEN_SYMBOL_ALLOW_ALL as the first thing in the .cpp file. Whenever
this is done, an explanatory comment should be added.
Note that this system cannot catch all "bad" usages (notably the Lua
code in the sword25 engine), as it can only work if scummsys.h is
included.
svn-id: r53961
Once the audio buffer has been queued in the audio stream, the Theora
decoder is no longer responsible for it. It has to allocate a new
audio buffer for the next set of audio data.
Of course, at the moment the audio is pretty useless because the
playback speed is all wrong.
svn-id: r53939
Even though our video player does not (yet) support audio, we need to
adjust the values of _audiobufFill and _audiobufReady as if it did, or
it will never realize it has reached the end of the cutscene.
Also fixes what looks like a typo in audio data clipping.
svn-id: r53938
