[TibixDev]

What

Add optional controller input through APS5_PAD_INPUT_FILE and tools/pad_input.py. Test runs can press buttons without desktop input injection. Pulses combine with physical input and release automatically.

Related: #1076 changes the input queue. This change adds a separate input source.

Tested

Fedora 44, Release: guest_pad passes. Tests cover pulse duration, stale and invalid commands, timestamps, triggers, and physical input. Python helper checks cover button commands, release, and invalid durations.

Checklist

• Current main; no duplicate open PR.
• One topic.
• No new code comments.
• Unsupported paths still throw; UI stubs are documented.
• No notes, logs, or images in the repository.
• No new third-party code.
• General behavior.
• Depends on: none.
• AI-assisted: yes.

[TibixDev]

Rebased on bc1fbe81. Kept the new remote-controller and special-port checks alongside file-input coverage.

Fedora 44 Release: guest_pad_tests build; the targeted tests and contribution checks pass. Windows CI is pending.

[adriadam10]

Checked on current main: build and the full ctest suite pass (guest_pad included), and CI is green on linux, windows and conventions. The code review found nothing blocking: the line, duration and button-mask bounds are enforced, invalid or stale commands are ignored, and the tests pin values.

@boykopovar this adds a new user-facing input interface (APS5_PAD_INPUT_FILE, tools/pad_input.py, a section in USAGE.md) that isn't covered by CONTRIBUTING, so it's a design call. Do you want a file-based injection channel in the emulator at all, or should test input go through another route? It also touches PadState.cpp, so it will need a rebase after #1076 and #1195.

[adriadam10]

Checked on current main after rebasing (one trivial conflict in GuestPad.cpp, both sets of checks kept): build and guest_pad pass, and the new test does not build without the change. The channel is opt-in (APS5_PAD_INPUT_FILE), bounds on line length, duration, button mask and sequence are enforced, and stale/invalid commands are ignored. Merging; whichever of this and #1076 lands second will need a rebase in PadState.cpp.

[blawesom]

Hi! This PR currently conflicts with main, so I rebased it onto the latest main (d3e84f85, includes #3296). The only conflict was in core/libs/tests/GuestPad.cpp, resolved by keeping both main's new scePadSetVibrationMode checks and this PR's CheckInjectedState(path) call. It builds and guest_pad_tests passes locally.

If it helps, the rebased branch is available to pull:

git fetch https://github.com/blawesom/AnyPS5 rebase/pr-1296
git reset --hard FETCH_HEAD

Feel free to ignore this if you'd rather rebase yourself.

[adriadam10]

Rebased onto current main after #1076 landed in PadState.cpp. The command-file buttons are now applied in both ReadState and Read (once per call, via a shared helper, to every returned state); the timestamp only changes when the pulse changes, as before. Build and 27 scoped tests pass (guest_pad included), and the new test does not build without the change. Merging once CI is green.