When running on Android, RetroArch considers most devices that emit dpad events as gamepads, even if they also emit other keyboard events; this is usually the right thing to do, but it has the side effect of not letting some actual keyboards (e.g.: Logitech K480) act as such inside RetroArch. This configuration option allows users to manually select a specific input device to act as a physical keyboard instead of a gamepad, which is handy when emulating computers as opposed to consoles.
- Fix analog drift blocking touch input (could occur on overlay_next if physical inputs shown on overlay)
- Fix overlay_next buttons lighting up in unison
- Skip meta keys in input_overlay_add_inputs (not supported by input_state_internal)
Allows stretching hitboxes and handling their overlap
reach_up, reach_down, reach_left, reach_right:
- Stretches in one direction:
reach_x, reach_y
- Stretches symmetrically
exclusive:
- If true, blocks input from overlapped hitboxes
range_mod_exclusive:
- Similar, but only applies when this hitbox is extended by range_mod
- After range_mod takes effect, has priority over 'exclusive'
E.g. This creates a D-Pad area and extends its hitbox left & right 50%, up 15%, and down 30%. Then applies range_mod_exclusive:
overlay0_desc0 = "dpad_area,0.15,0.57,rect,0.166228,0.295516"
overlay0_desc0_reach_x = 1.5
overlay0_desc0_reach_up = 1.15
overlay0_desc0_reach_down = 1.3
overlay0_desc0_range_mod = 2.0
overlay0_desc0_range_mod_exclusive = true
* Added usb hid controllers for the famous ZeroDelay encoder and also for "Kade: Kick Ass Dynamic Encoder" to be able to use some custom arcade sticks.
* Add accidentally missing change to Makefile
Co-authored-by: DiktatorShadaloo <diktatorshadaloo@gmail.com>
- New descriptors: 'dpad_area' and 'abxy_area'
- Each has a diagonal sensitivity setting, 100% being 8-way symmetry
- Buttons can be redefined in the cfg file
E.g. This would create a d-pad area, then redefine it to use analog directions:
overlay0_desc0 = "dpad_area,0.85,0.57,rect,0.166228,0.295516"
overlay0_desc0_up = "r_y_minus"
overlay0_desc0_down = "r_y_plus"
overlay0_desc0_left = "r_x_minus"
overlay0_desc0_right = "r_x_plus"
Repurpose vibrate_on_keypress to enable device's standard keypress feedback on overlay key/button state changes
- Add keypress_vibrate function ptr to input_driver_t (only implemented on Android for now)
- (Android) Remove APP_CMD_VIBRATE_KEYPRESS
- (Android) Add doHapticFeedback, called directly to avoid latency
* * Use flags for rarch_state
* Get rid of ^M linebreaks in retroarch_types.h
* Buildfixes for consoles
* (audio driver) use flags instead of bools
* (video) Use flags instead of bools
* Rewrite input driver state bools into flags
* Fix Sixaxis gamepad operation
== DETAILS
It looks like commit 78e7d23c14 broke
the DualShock. Sorry, @revvv, but this is probably going to break the
Retrode driver. We'll need to figure out a different solution.
* Fix a bunch of HID implementations
== DETAILS
As mentioned in the previous commit, the thing that broke the DS3
driver was a change to the wiiu HID handler, which did some pointer
math on the data buffer.
The thing is.. there's no reason to do that pointer math. Yet, I found
the same thing on almost every other HID driver: OSX, libusb, wiiusb.
The only other HID driver that did not do this was the Bluetooth
HID driver.
It doesn't make any sense. The only reason it doesn't crash is because
the memory right before the buffer is valid memory.
Change summary:
- remove the weird pointer math from all HID controllers
- remove the `#ifdef apple ...` in the ds3 and gamecube adapter
drivers, because it's no longer needed.
- in the pad packet handlers, decrement data buffer offset references
to account for the removal of the buffer manipulation.
* Fix DualShock 4
== DETAILS
Should've only subtracted 1, not 2; and now the pad works.
Also, added a PID to cover newer model DS4s. I picked "R2" out of the air
for the constant.
Tested on Mac OS
* Really really fix iohidmanager_set_report
A huge apology to @23rd for insulting your fix. I was wrong to call you
wrong.
That left the question: why did that change break the DS3?
Well, it's because `IOHIDDeviceSetReport` expects the report buffer to be
just the report. All of RA's HID SetReport calls include the report number
in the first byte, so we have to trim that off.
We actually do this in a different invocation of `IOHIDDeviceSetReport`
elsewhere in the file! This commit applies that same logic to
`iohidmanager_set_report`
This has two benefits:
1. The DS3 works
2. The DS3 no longer requres the user to press the PS button to activate
it the first time. You plug it in and it Just Works, just like on Wii U.
* MacOS: Fix sixaxis on MacOS 12.x
== DETAILS
Mac OS 12.x "Monterey" dropped some built-in support for the
Sixaxis controller--exactly what changed is unclear, but other
projects reported similar issues with the OS.
This commit ports the work done on SDL, which is basically to
read two feature reports after writing out the activation report.
I think we did this at one point in the past, but it got lost
among HID driver rewrites.
I've tested this locally and it resolves the issue; after connecting
the sixaxis and pressing the Playstation button, the pad is registered
and works successfully.
I have not tested this on other platforms.
* Gracefully handle incomplete HID implementations
== DETAILS
Not all HID drivers implement all HID methods; since this controller
has a magic sequence that requires the ability to set and get HID
reports, we need to make sure we don't crash if we're running on
an incomplete implementation.
- Add a null check for the driver methods and log useful info if checks
fail
- Fix minor formatting
* Reduce snprintf and/or use them only for concatenating the
string formatting of numbers/values
* Reduce snprintfs
* Use snprintf for concatenation at parts
* * We don't need to NULL-terminate strings that get passed to strlcpy/strlcat
and friends
* Use snprintf for concatenation in certain instances
* Do away with some string intermediary copies where we can avoid it
* Fix warning unused variable
* * Reduce snprintf calls
* Rewrite snprintf calls into strlcpy where possible
* Use snprintf for concatenation when necessary
* Do away with some string intermediary copies in task_translation.c
* run_translation_service - make switch slightly smaller
(snprintf) Try to reduce or simplify snprintf calls, only tend to
use it for processing integers/numbers and avoid it for regular
string concatenation (NOTE: we try to be a bit safer about it to
address earlier cited criticism, although we don't consider concatenating
3 or 4 characters at the end to be insecure)
(msg_hash_to_str) Try to avoid duplicate calls to the same localized
string when we can just cache the results once instead locally
- (database_info.c) Optimizations for database_info_build_query_enum() -
it would call strlcat at the very least 3 times, normally more than that.
Every strlcat call is at least one or more strlens. Reduce it all down
to manual string concatenation and one strlcat
- (database_info_build_query_enum) - always NULL terminates string so no need for NULL termination done by callee
* (menu_displaylist) use smaller strings for basename paths, remove one unnecessary string
in menu_displaylist.c function
- (disk_index_file.c) Avoid another trivial string concatenation by doing
it manually after fill_pathname_join_special
- (fill_pathname_join_delim) since fill_pathname_join_delim() now has a return value, use it for
string concatenation
- (platform_unix) - Use BUFSIZ as size for a buffer in system_property_get
- some other trivial strlcat usage replacements
as deprecated.
* Use fill_pathname_join_special in the vast majority of cases where
we can ensure out_path is a new empty string
* Get rid of some extension concatenation with strlcat where encountered
* Some general cleanups with NULL termination of strings that get immediately
passed to strlcpy/strlcpy-adjacent functions
* (joypad_connection) Small optimizations -
* Turn functions static where possible
* Hose strlen call out of loop
* (input_driver.c) General cleanups:
* Some small code/style nits
task_screenshot.c:
* Move widget callback function for screenshots to task_screenshot.c
(file_path.c):
* Turn get_pathname_num_slashes into static function
* path_linked_list_free - always returns true, so get rid of return value
* path_linked_list_new - fix function signature
* path_get_archive_delim - do not NULL-terminate string, already done by strlcpy later on
General:
* Slight optimizations - use int/size_t for loop counter variable instead of unsigned
* Take advantage of fact that strlcpy already NULL-terminates, so don't do this explicitly
outside if we're just going to end up calling strlcpy/fill_pathname_join on it anyway
Closes https://github.com/libretro/RetroArch/issues/3414
I have investigated the issue. The crux of the problem is that on Android there
is no way distinguishing 2 scenarios:
1) 2 identical bluetooth controllers A and B and first there are button presses
only on controller A and then on controller B
2) the same controller disconnects and reconnects.
Android doesn't give bluetooth mac address of where the touch came from, only
opaque ID and this opaque ID changes after reconnect. Hence without changes to
android this is infeasible without giving up the ability for 2 users to play on
identical controllers.
I guess that this sacrifice makes sense for affected users
Controllers with same VID/PID are distinguished by the device name.
1. The Wii U only sends a prefix of the device name.
2. The check preferred the device name over VID/PID which was not
intended.
Example: The device name "USB Gamepad" is truncated to "USB" which
was mapped to "Generic SNES USB Controller", although VID/PID did not
match.
Otherwise autoconfig file can't be found on the Wii U:
VID/PID are not available, so it has to match with 'input_device' and I
don't want to add additional autoconfig files with different
'input_device' strings.
Otherwise the USB gamepad cannot be found, if VID/PID has leading zero.
For example: Retrode gamepad adapter:
vid=1027 (0x403) pid=38849 (0x97c1) becomes
vid=1027 (0x403) pid=-26687 (0xffff97c1)
* features_cpu: Fix clock for djgpp
* frontend_dos_set_fork: Fix wrong id names
* dos_input: Fix missing indexing
* dos_joypad: Fix wrong id name
* Fix menu drawing
Menu is already prescaled to VGA_WIDTH x VGA_HEIGHT in vga_menu_frame. We need
to always use those dimensions
* Change dos compilation from griffin to common
* Set vga menu size to 320x200
* Add DOS compilation in gitlab-ci.yml
* Support for Swift, added emulator keyboard
* fixed toggle key handling using magic number hack for now
* fixed keyboard transparency slider for now with suboptimal fix; add gesture recognizer to hide/show keyboard
* Support CocoaView extensions in Swift; move keyboard delegate impl and setup to swift extension
* moved keyboard view model creation out of EmulatorKeyboard
* implement key pressed delegate in swift extension
* added input method for directly sending RETROK_* codes to support a touchscreen keyboard; assign keyboard model delegates; updated keyboard layout (added F1-F12 keys); change shift, control and alt keys to be modifiers
* enable focus mode when custom keyboard is shown; enable/disable overlay when custom keyboard is toggled
* Specify -DHAVE_OPENGLES2 instead of -DHAVE_OPENGLES3 since glsym_es3.h does not compile in iOS 14.5
Fix tvOS build using compiler flags
* Create new project for iOS 13 deploy target; add check for deploy target to conditionally compile for new iOS 13 specific feature (custom keyboard)
* force disable core info caching for iOS, use opengl es2 for debug
* Add flag for iOS custom keyboard - 3-finger swipe up to show, 3-finger swipe down to hide
* use OpenGLES2 instead; using ES3 results in compile time errors on iOS 14.5
* code cleanup
* Updated references to -DDONT_WANT_ARM_ASM_OPTIMIZATIONS flag
* Add JIT support for non-jailbroken devices
* iOS: Add support for touch mouse handler
* Added a HAVE_IOS_TOUCHMOUSE preprocessor macro so that it builds under the iOS11_Metal xcode project
* Changed click-and-drag behavior to double tap hold and drag
* Visual improvements to the emulator keyboard: updated colors, improved key-press effect
* (Wayland) Show window early to get screen info
* Report current monitor size on initial get_video_size call
* Fullscreen to monitor selected in settings
* Fullscreen on current monitor when setting is set to automatic
* Fallback to shm_open when memfd_create is not supported
* Report current monitor size on initial get_video_size call
* Fullscreen to monitor selected in settings
* Fullscreen on current monitor when setting is set to automatic
* Added newlib changes
* Add action to launch PPSSPP simulator
* Remove legacy config for the stack and heap
* Add pthread
* Simplify kernel_functions and improve references to HAVE_KERNEL_PRX
* Add some flags
* Improve audio init/deinit
* Improve exit by clicking home
* Add CI for PSP1
* Update PSP.yml
* Fixes mouse-grab toggling in Wayland.
Fixes#13536
* Indicate that input_wl_grab_mouse is a necessary function.
It might prevent it from being removed again in future cleanups.
* Netplay Stuff
## PROTOCOL FALLBACK
In order to support older clients a protocol fallback system was introduced.
The host will no longer send its header automatically after a TCP connection is established, instead, it awaits for the client to send his before determining which protocol this connection is going to operate on.
Netplay has now two protocols, a low protocol and a high protocol; the low protocol is the minimum protocol it supports, while the high protocol is the highest protocol it can operate on.
To fully support older clients, a hack was necessary: sending the high protocol in the unused client's header salt field, while keeping the protocol field to the low protocol. Without this hack we would only be able to support older clients if a newer client was the host.
Any future system can make use of this system by checking connection->netplay_protocol, which is available for both the client and host.
## NETPLAY CHAT
Starting with protocol 6, netplay chat is available through the new NETPLAY_CMD_PLAYER_CHAT command.
Limitations of the command code, which causes a disconnection on unknown commands, makes this system not possible on protocol 5.
Protocol 5 connections can neither send nor receive chat, but other netplay operations are unaffected.
Clients send chat as a string to the server, and it's the server's sole responsability to relay chat messages.
As of now, sending chat uses RetroArch's input menu, while the display of on-screen chat uses a widget overlay and RetroArch's notifications as a fallback.
If a new overlay and/or input system is desired, no backwards compatibility changes need to be made.
Only clients in playing mode (as opposed to spectating mode) can send and receive chat.
## SETTINGS SHARING
Some settings are better used when both host and clients share the same configuration.
As of protocol 6, the following settings will be shared from host to clients (without altering a client's configuration file): input latency frames and allow pausing.
## NETPLAY TUNNEL/MITM
With the current MITM system being defunct (at least as of 1.9.X), a new system was in order to solve most if not all of the problems with the current system.
This new system uses a tunneling approach, which is similar to most VPN and tunneling services around.
Tunnel commands:
RATS[unique id] (RetroArch Tunnel Session) - 16 bytes -> When this command is sent with a zeroed unique id, the tunnel server interprets this as a netplay host wanting to create a new session, in this case, the same command is returned to the host, but now with its unique session id. When a client needs to connect to a host, this command is sent with the unique session id of the host, causing the tunnel server to send a RATL command to the host.
RATL[unique id] (RetroArch Tunnel Link) - 16 bytes -> The tunnel server sends this command to the host when a client wants to connect to the host. Once the host receives this command, it establishes a new connection to the tunnel server, sending this command together with the client's unique id through this new connection, causing the tunnel server to link this connection to the connection of the client.
RATP (RetroArch Tunnel Ping) - 4 bytes -> The tunnel server sends this command to verify that the host, whom the session belongs to, is still around. The host replies with the same command. A session is closed if the tunnel server can not verify that the host is alive.
Operations:
Host -> Instead of listening and accepting connections, it connects to the tunnel server, requests a new session and then monitor this connection for new linking requests. Once a request is received, it establishes a new connection to the tunnel server for linking with a client. The tunnel server's address and port are obtained by querying the lobby server. The host will publish its session id together with the rest of its info to the lobby server.
Client -> It connects to the tunnel server and then sends the session id of the host it wants to connect to. A host's session id is obtained from the json data sent by the lobby server.
Improvements (from current MITM system):
No longer a risk of TCP port exhaustion; we only use one port now at the tunnel server.
Very little cpu usage. About 95% net I/O bound now.
Future backwards compatible with any and all changes to netplay as it no longer runs any netplay logic at MITM servers.
No longer operates the host in client mode, which was a source of many of the current problems.
Cleaner and more maintainable system and code.
Notable functions:
netplay_mitm_query -> Grabs the tunnel's address and port from the lobby server.
init_tcp_socket -> Handles the creation and operation mode of the TCP socket based on whether it's host, host+MITM or client.
handle_mitm_connection -> Creates and completes linking connections and replies to ping commands (only 1 of each per call to not affect performance).
## MISC
Ping Limiter: If a client's estimated latency to the server is higher than this value, connection will be dropped just before finishing the netplay handshake.
Ping Counter: A ping counter (similar to the FPS one) can be shown in the bottom right corner of the screen, if you are connected to a host.
LAN Discovery: Refactored and moved to its own "Refresh Netplay LAN List" button.
## FIXES
Many minor fixes to the current netplay implementation are also included.
* Remove NETPLAY_TEST_BUILD
* convert abs mouse from screen to viewport coordinates; fix relative mouse code to work in screen mode
* C89 compatibility
* revert accidental include
* Allow parallel compilation in PS2
* Allow to compile with griffin or common compilation in PS2
* Enable dummy core to be used in other platforms
* Use threads in YML config
* Add the compilation to PS2 in GitHub Actions
== DETAILS
So, the reason the gamepad was getting deregistered was
because adapter free code wasn't properly handling null-interface
adapters, causing the gamepad to match erroneously and get
deregistered.
This doesn't fix the weird "Generic SNES USB" detection issue,
but it should make it non-fatal.
Co-authored-by: Nathan Strong <nstrong@tripwire.com>
== DETAILS
The problem was caused by changing the driver lookup point. The deferred
lookup resulted in the WIIU trying to treat the keyboard as a gamepad which ...
didn't work.
This change short-circuits at the connection event by ignoring mouse &
keyboard connection events.
Tested this with a dev build.
Co-authored-by: Nathan Strong <nstrong@tripwire.com>
* Any pad can control the menu
== DETAILS
I am not sure I've quite got it so that any pad can *open* the
menu, but I do have it so any pad can control it.
- split out the input processing into a separate method
- track down and squish some hairy bugs that boiled down to
bad pointer math
- it looks like `menu_driver.c` has a mix of line endings, so I
ran it through `dos2unix` so it has consistent line endings
again.
- verified that this change did not impact actual cores
* optimize out cumulative_bits
* Incorporate PR feedback
Many thanks to @jdgleaver for providing these optimizations.
* apply one more optimization
* Move more state to runloop state
* remove unused variable
* Cleanup
* Move more state to runloop_state
* Remove unused variable
* Cleanups
* move input_remapping functions over to input_driver.c
* Some buildfixes
== DETAILS
The crashes he reported in the gbatemp thread are due to iface
being null (since it's an unsupported device) and unchecked
iface dereferences.
== DETAILS
- only call disconnect when we're actually disconnecting a remote
(e.g. read errors or remote goes to sleep).
- clean up some compile warnings introduced by others (mainly
unused variables)
== DETAILS
Put the finishing touches on getting the DS3 to work on Mac OSX.
Basically, there's some differences in the HID interface bewtween
wiiu and osx where OSX expects the first byte of the report to be
the report ID, while wiiu expects that byte to be trimmed off.
I was able to put this behavior in the respective HID
implementations, which eliminated the confusing packet offset
ifdefs.
And, I was able to get the LEDs working again.
== DETAILS
The DS3 driver previously only worked with the Wii U HID implementation.
I adapted this driver from the Linux driver for the DS3. It's not quite
100%--I haven't got the LEDs to work properly--but it's functional.
Going to continue tweaking it to see if I can get the LEDs to work.
== DETAILS
- rewrote the HID deregistration algorithm; it should no longer
cause issues when dealing with multiple pads of the same HID/VID
combo
- fix initialization bug that caused wiimotes to fail to register
without an accessory attached
== DETAILS
The GCA uses a weird HID class that our current filters don't
catch, so we add it.
Needed to do a small amount of tweaking on the GCA driver to
account for iohidmanager weirdness.
== DETAILS
The HID drivers in `input/connect/` were missing an implementation
for button(), so I added it. The only exception is the wii driver,
which is really complicated and more than I wanted to try to tackle--
especially since WIIU has its own wiimote drivers.
== DETAILS
Use a little trickery to ensure the GCA driver continues working
with other HID implementation.
I've expanded the joypad implementation to support multi-pad devices.
However, this requires changes to each HID implementation to actually
function.
I've made the necessary changes for WIIU, but I don't have the means
of making the change in the other HID implementations.
So, I've built in a backwards-compatibilty mode for the driver.
The trick is to have an identifier byte at the top of both data structs
that the driver returns. We can then use that byte to determine which
of the structs has been passed to the pad functions and act accordingly.
In the GCA case, for non-wiiu platforms, it will simply expose port 1
of the GCA and the other 3 ports do nothing.
== DETAILS
cause of crash: trying to deference init when it's null
the reason it was going into deregister: the HID/VID lookup was
failing because it wasn't getting initialized first
cause of 2nd crash: the "end of pad list" method looked for an entry
with a magic value, so we add an end marker
== DETAILS
I'm going to need to re-implement the abstraction between HID device
and gamepad; the changes here lay down some foundation for that work.
- reduce logging priority of some statements I modified while debugging
- factor out the pad driver lookup into its own method
- fix so the pad driver list isn't re-initialized every invocation
- add the button() method for the GCA driver
== DETAILS
When I first implemented the Wii U HID architecture, I ended up
having to design my own implementation because, at the time, I did
not have a way to read the HID device string to allow the existing
code to successfully detect the gamepad.
After spending some time experimenting, I've figured out how to
do this. And that means I can better align the HID driver with other
platforms.
change summary:
- create a single state structure for all three sub-types of wiiu pads
(kpad, wpad, and hid)
- eliminate confusing duplicate pad lists
- eliminate confusing duplicate HID pad drivers (ds3, gamecube
adapter, etc)
- ensure the ds3 driver still works
sonninnos