When loading and saving states taken during a replay, their contents
should be compared with the current replay.
- If in playback mode, states from past the end of the current replay
should be disallowed, as should states from an alternate timeline of
the current replay (inconsistent input history)
- This is because in playback mode, RA replays are read-only
- If in recording mode, states from the future or from another
timeline should overwrite the current replay's data with a warning
- In either playback or recording, loads of states from the past that
have a consistent series of inputs (same timeline) should be loaded
and the replay should be rewound (and truncated in case of recording)
This behavior is as specified in
https://tasvideos.org/LawsOfTAS/OnSavestates
The existing code only implemented some of these checks.
This can optionally encode and compress checkpoints. For now, only
raw encoding is supported, but e.g. delta encoding, block-based
encoding, or deduplicated encoding could be supported in the future.
Three compression schemes are included (selected automatically based
on compilation flags): ZSTD, ZLIB, or NONE.
All three schemes are supported for decompression if available.
For compression, ZSTD is preferred over ZLIB is preferred over NONE.
- rewinding past the start clobbers part of the header
- during recording & rewind, should not push input events
- during recording or playback, getting to the end of the rewind
buffer should not halt the movie
Check network fd/socket validity using INVALID_SOCKET on Windows
(requiring conditional winsock2.h include) and < 0 otherwise.
Fixes -Wsign-compare warning.
- Ensure other touchscreen input is blocked when overlay buttons pressed
- Limit physical+overlay mouse input to iOS for now
- Reset overlay mouse deltas when there is no touch input
- Disable touch mouse in android_input.c for now. Too buggy
* Actually read CLI args in emscripten
* Fix fetchfs manifest parsing, increase download chunk size
The chunk size should probably be made a parameter in the future. The
larger chunk size trades longer hitches for fewer hitches.
* Add exec command driver and API functions for emscripten.
Under WASMFS, stdin/stdout can't be customized the way they can with
the JS FS implementation. Also, this approach frees up stdin/stdout
and simplifies interaction with the command interface for web embedders.
* fixup upload paths, show use of new emscripten cmd interface
* Add JS library function names to EXPORTS as well as EXPORTED_FUNCTIONS for older emsdk versions
* workerized RA
* Workerized (non-async) web player, using OPFS
This patch eliminates the need for asyncify and uses modern filesystem
APIs instead of the deprecated, unmaintained BrowserFS.
This is a WIP patch because it won't fully work until these two
Emscripten PRs land and are released:
https://github.com/emscripten-core/emscripten/pull/23518https://github.com/emscripten-core/emscripten/pull/23021
The former fixes an offscreen canvas context recreation bug, and the
latter adds an equivalent to BrowserFS's XHR filesystem (but without
the hazardous running-XHR-on-the-main-thread problem).
The biggest issue is that local storage of users who were using the
old version of the webplayer will be gone when they switch to the new
webplayer. I don't have a good story for converting the old BrowserFS
IDBFS contents into the new OPFS filesystem (the move is worth doing
because OPFS supports seeking and reading only bits of a file, and
because BrowserFS is dead).
I've kept around the old libretro webplayer under
pkg/emscripten/libretro-classic, and with these make flags you can
build a non-workerized RA that uses asyncify to sleep as before:
make -f Makefile.emscripten libretro=$CORE HAVE_WORKER=0 HAVE_WASMFS=0 PTHREAD=0 HAVE_AL=1
I also moved the default directory for core content on emscripten to
not be a subdirectory of the local filesystem mount, because it's
confusing to have a subdirectory that's lazily fetched and not
mirrored to the local storage. I think it won't impact existing users
of the classic web player because they already have a retroarch.cfg in
place.
* Get fetchfs working without manifest support
* makefile fixes
* fix scaling, remove zip dependency
* Support asset/cheats/etc downloaders for emscripten
- Add http transfer support for emscripten
- At the task_http level, not the net_http level --- so no netplay
or webdav.
- Change default paths to be more like other platforms
- Gives us smaller bundles and a faster boot time
- Had to work around a task queue bug on Emscripten
- I made the smallest possible change to do it, but it may be better
to fix in rthread.c
* Load an emscripten file_packager package on first run
If no ozone assets are present, load a libretro_minimal package
created using Emscripten's built-in file packager.
* updated readme, removed indexer from wasmfs libretro-web
* Put back zip dependency, load asset bundle into opfs on first run
* fix upload path
* Remove unused function
* easy testing setup for two multithreaded conditions
1. make PROXY_TO_PTHREAD=1 (slower)
2. make PROXY_TO_PTHREAD=0 (bad audio, because doesn't sleep in
openal.c)
* Remove condition on sleep in openal
also make input_driver check existence of drv->axis, drv->button
before calling them.
* Fix resizing under EGL
* Don't force config file path on emscripten
* Add time.h include to netplay, default HAVE_NETPLAYDISCOVERY to 0
* Remove nearly all proxied joypad calls under emscripten
* Fix file uploads under firefox
* Fix safari API uses, but Safari still hangs in OPFS filesystem mount
I think this can be fixed by moving the backend creation off the main
thread.
* Move filesystem init into emscripten C entry point
* Setup filesystems off of main thread
* re-set default player to async
Also improve Safari compatibility under proxy-to-pthread condition
* Safari upload file fixes
* Remove some excess prints
* Fix typo
* Add dummy game ai subsystem
* First working prototype of a machine learning model that can override player input
* Update README.md
* Update README.md
* Fix loading path on Windows
* Change ai override to player 2
* Added quick menu show game ai option
* Implemented Quick Menu entry for Game AI options
* Redirect debug logs to retroarch log system + properly support player override
* Added support to use framebuffer as input to the AI
* Added pixel format parameter to API
* Fix game name
* code clean-up of game_ai.cpp
* Update README.md - Windows Build
* Update README.md
* Update README.md
* Update README.md
* Update config.params.sh
turn off GAME_AI feature by default
* Fix compile error in menu_displaylist.c
* Add missing #define in menu_cbs_title.c
* Added new game_ai entry in griffin_cpp
* Remove GAME_AI entry in msg_hash_us.c
* Fix compile error in menu_displaylist.h
* Removed GAME AI references from README.md
* Fixes coding style + add GameAI lib API header
* Convert comment to legacy + remove unused code
* Additional coding style fixes to game_ai.cpp
* Fix identation issues in game_ai.cpp
* Removed some debug code in game_ai.cpp
* Add game_ai_lib in deps
* Replace assert with retro_assert
* Update Makefile.common
* Converting game_ai from cpp to c. First step.
* Convert game_ai from CPP to C. STEP 2: add C function calls
* Convert game_ai from CPP to C. Final Step
* Added shutdown function for game ai lib
* Update game_ai_lib README
* Fix crash when loading/unloading multiple games
- Fix the "Allow Off-Screen" setting not working since #17308
- Prevent offscreen shot from disabling the lightgun with some cores
- Fullscreen pointer coordinates no longer needed
Adapt the sanitized pointer handling, discussed at #17196 :
Overlay "driver" specific changes:
- make sure pointer position is always within [-0x7fff,0x7fff] by using the confined wrapper
- enable pointer offscreen query
- report -0x8000 for lightgun if pointer is at the edge
- align lightgun offscreen reporting and button ID conversion with other drivers
Android driver specific changes:
- make sure pointer position is always within [-0x7fff,0x7fff] by using the confined wrapper
- remove extra "inside" checks, general simplification
- enable pointer offscreen reporting
- report same value for all ports when querying mouse and lightgun
- fill missing lightgun support, with fixed button map
Udev and X11 driver specific changes:
- simulate max. 3 touches instead of 1 using different mouse buttons
Wayland driver specific changes:
- integrate touch input better to the overall handling (enabling overlay usage with mouse)
- simulate max. 3 touches instead of 1 using different mouse buttons
* Less string copies
* Some general cleanups
* Add extra param to runloop_message_queue_push so we can pass size_t
of the message
* Consistent conventions for local variable usage for certain things
Adapt the sanitized pointer handling, discussed at #17196 :
X11 driver specific changes:
make sure pointer position is always within [-0x7fff,0x7fff] by using the confined wrapper
remove extra "inside" checks, general simplification
enable pointer offscreen reporting
Udev driver specific changes:
remove custom calculation and use common viewport translation
unify pointer query instead of separate _x and _y
enable pointer offscreen reporting
Other changes:
more tuning of pointer conversion in video_driver.c for edges
lightgun button ID conversion moved to input_driver.c
Adapt the sanitized pointer handling, discussed at #17196:
Wayland driver specific changes:
- make sure pointer position is always within [-0x7fff,0x7fff]
by using the confined wrapper
- enable lightgun to report -0x8000 if pointer is really offscreen
- remove extra "inside" checks
- report same pointer/lightgun coordinates for all ports
- simplify pointer and lightgun handling
Other changes:
- unify "offscreen" condition through input_driver.c
- slight tuning of pointer conversion in video_driver.c
- update libretro.h with explanation and pointer offscreen value
- small fixes on remote retropad test screen
Changed the logic of the classic toggle turbo mode code such that the pressing order of the turbo button and the button to enable/disable does not matter.
libretroadmin