The controller sound card path already scales its samples by ANYPS5_RUMBLE, but
the Bluetooth haptics stream took the vibration port samples before that, so
ANYPS5_RUMBLE=0 still left a Bluetooth DualSense vibrating.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The worker sent a backlog of buffered blocks back to back. It now sends at
most one report 0x32 per 10.67 ms, the length of one 32-frame block at
3 kHz, without a catch-up burst after a pause.
A failed report write cleared the stream-open flag, which sent the vibration
ports to motor rumble. The flag now stays set until the controller
disconnects, so the ports stay silent while the stream is stopped.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The 10.672 ms timer took a block only when it fired, so audio arriving in
blocks smaller than 32 haptic frames (16 from a 256-frame legacy port)
regularly found fewer than 32 frames and the report was skipped; catching
up after falling behind sent two reports back to back and left a ~21 ms
hole. The controller, asked to queue only 16 frames, then ran dry, and
the haptics dropped out about once per second on hardware.
The stream thread now waits on a condition variable until two blocks are
buffered and sends one, so reports follow the audio clock, and report
0x32 asks the controller to queue 64 frames (21 ms) to absorb Bluetooth
jitter and the difference between the title's rate and the controller's
crystal. Closing the stream wakes the waiting thread before the join.
Co-authored-by: seradavid <[email protected]>
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Recorder keeps both this branch's LookupStoredBytes and main's UnlandedOwnLabel, with both test groups in --completion-labels-only and the full recorder run.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Six flips under ANYPS5_FPS_LIMIT=20 may not complete in under five periods
of 50 ms. The check is a lower bound only, so a loaded host cannot fail it;
without the hook the six flips complete in about 100 ms at the output's
rate.
main moved the AGC test registrations into tests/cmake; tools/agc_test_split.py resolve moved this branch's lines to their fragments.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
setenv is not part of the MinGW Windows CRT. The test sets the variable
for the process, which PadIdentify.hpp reads with std::getenv. SDL_setenv
calls SetEnvironmentVariableA on Windows, which does not update the CRT
table that std::getenv reads, so _putenv_s is used there. Linux keeps
setenv.
Four process-level checks on an SDL virtual controller whose rumble
callback records the motor edges: connect sends exactly one pulse for
player 1 after the quiet window; off sends nothing; a title rumbling
at 200 gets no identify pulse while it rumbles and one pulse after it
stops; an invalid ANYPS5_PAD_IDENTIFY stops the start with an error
naming the variable and the value. The test skips with code 77 when the
host has joysticks, as pad_sdl does.
A scripted button that answered a dialog stayed masked from the title
until a title read sampled it released. When the title did not read
during the release, a later scripted press of the same button was
masked as well. The mask now keeps only the buttons the command
timeline held without a break since the last read, so a press that
ends, or a new command, clears it whether or not it was sampled.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The removal and reconnect checks sampled only at the hold's edge, so a hold
of zero still passed them. A sample halfway through the hold must now stay
pending, which a zero hold does not do.
A scripted press that answers a modal dialog or the IME was merged into
the host state for the capture, but video out's heldThroughCapture only
sees host buttons, so once the capture ended Pad::ReadState OR-ed the
still-active injected press into the title's read. A 400 ms Cross that
answered the dialog then also reached the title for the rest of the
press.
Record port 0's injected buttons while the pad is captured and mask them
from the title until they are released, as host buttons already are.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Require aborted without saying which check failed, so a failure on a CI host that cannot be reproduced locally left no trace of the assertion in the log. It now prints the source line before aborting.
Nothing happens by default, focus lost silences and focus gained restores exactly unity with the option on, other window events, other windows and key events are ignored, and the user's own mute survives a focus round trip.
Register pad_host_volume. An empty request set changes nothing, volume up and down move one step and return to exactly unity, mute toggles, combined requests run up then down then mute, and the other host requests are ignored.
No built-in key is bound twice, the three host volume keys are bound by default as plain keyboard keys, a key the user shares with another action keeps both bindings as it does for the existing actions, and action names are exact and case-insensitive.
It takes a key and alternates case-insensitively, ignores a duplicate line, and refuses mouse buttons, the wheel, unknown keys, empty sources and sources without a type.
It takes a key and alternates case-insensitively, ignores a duplicate line, and refuses mouse buttons, the wheel, unknown keys, empty sources and sources without a type.
It takes a key and alternates case-insensitively, ignores a duplicate line, and refuses mouse buttons, the wheel, unknown keys, empty sources and sources without a type.
Runs the default mode first (port 0 and user 0 stay connected after the only
controller is removed), then controller mode with virtual SDL controllers:
the removal is committed after the hold, user 0 logs out and back in, a
withdrawn port 0 reads zeroed, and a brief removal does not flap port 0.
