Sylpheed/docs/port/held-direction-repeat.md

F1 — the menu repeats on a held direction: mechanism shipped, and the rate adopted

Status: ✅ mechanism implemented and wired, 2026-09-02. ✅ rate adopted 2026-09-13 from docs/re/f1-repeat-measured-via-driver-patch.md — REPEAT_DELAY = 0.402, REPEAT_INTERVAL = 0.134. It ran inert for eleven days and that was the right state; this page keeps the inert-era reasoning because it is why the two numbers can be trusted now.

What was reported

Two statements from the human, both about the real game, a play-test apart:

"Moving stick up/down and holding only moves one item. In game it actually continues to move when holding up/down, just at a medium pace so player does not need to move pad middle↔up/down, but also slow enough to see which item is selected and move to target."

"Confirmed D-Pad does repeat when holding too."

The file predicted its own refutation

gamepad.gd carried this, written when the latch was added:

"Whether the real game repeats while a direction is held, and how fast, is unknown… If the game does repeat, this is a difference a human will notice as 'I have to flick it again', and the fix is a measured repeat interval — not a guessed one."

That is exactly what happened, in the words it predicted. So one-step-per-deflection is no longer the conservative reading — it is a known defect, and keeping it is choosing a wrong behaviour over an approximate one.

What was built

Gamepad.held_direction()	−1 / 0 / +1, polled from the devices
Gamepad.repeat_due(delta)	one step or 0, per frame
Boot._menu_repeat(delta)	calls it under the same guards a real press gets

Why it polls devices and not Input.is_action_pressed

ui_up/ui_down are bound to the stick axis at Godot's 0.50 action deadzone, while this port steps at the game's measured 0.61 (ENTER). Polling the action would repeat throughout the 0.50–0.61 band — the exact band ENTER exists to exclude — so the repeat would contradict the threshold on the same stick, on the same frame.

That is the input-map lesson from 2026-09-01 arriving in a new place: assert the device, not the layer above it. The stick reads from the latch accepts() already maintains, so the first step and the repeat cannot disagree about hysteresis; the d-pad reads JOY_BUTTON_DPAD_UP/DOWN directly, which the human's second report makes load-bearing rather than defensive.

Why the guards are duplicated rather than shared

_menu_repeat re-applies the same four conditions _unhandled_input applies — no movie playing, a menu exists, its stack is non-empty, no transition pending. A repeat that could fire during a movie or mid-transition would be a second, subtly different input path, and the first thing this port learned about input is that a second path is where the defect hides.

✅ The rate, and how far each half of it reaches

An earlier draft had REPEAT_DELAY = 0.40 and REPEAT_INTERVAL = 0.20 with a paragraph explaining that they were authored. They were removed rather than commented out, on an explicit instruction:

"Take the RATE from the Decoder — do NOT ship a placeholder interval. An invented rate here is indistinguishable from a measured one later, and this is the exact field where that already cost us."

📌 The measurement has now landed, and it vindicates the instruction in the most awkward possible way: the guessed delay was nearly right and the guessed interval was off by 50 %. 0.40 against a measured 0.402; 0.20 against a measured 0.134. Had both shipped, the half that was wrong would have been protected by the half that was right.

The Decoder measured, in Canary at an achieved 29.87 fps guest rate, 12 frames from the press-triggered step to the first repeat and 4 frames per step after it. Converted to seconds here because this port does not run at the guest's rate and it is the cadence that was measured: 12 / 29.87 = 0.402, 4 / 29.87 = 0.134.

⚠️ The two numbers are not equally well evidenced

No physical controller exists in the Decoder's container. The measurement was taken by patching Canary's --hid=file driver to emit Keystroke REPEAT at the SDL driver's own 400 ms / 100 ms constants:

• the delay came back as 402 ms — to within the frame quantum, the constant that was fed in. It confirms the instrument, not the game;
• the interval came back as 133 ms against a fed-in 100 ms. That gap is the genuinely new fact: the game paces repeats to its own frame consumption rather than to the event stream.

So the interval is what the game does; the delay is what Xenia's SDL driver does, and the game was not observed to disagree. One run — the corpus's own two-run minimum is not met, and the source page says so itself.

One thing about the rate was already measured, and it narrowed the question. The game digitises the left stick to four direction bits at 61 % deflection, so it cannot see deflection magnitude at all — the repeat it drives cannot be faster-the-harder-you-push. That excluded the one competing model, so only two constants were ever open and a single measurement closed both.

🔴 Adopting the rate was predicted to break a green check. It did not, and that was worse

This page and gamepad.gd both said that verify-input's row "a held stick is ONE step, not six" asserts the absence of the repeat, and would go red on adoption — looking like the 2026-09-01 jitter defect returning.

Run on adoption day: the row stayed green. steps() feeds axis values through the latch and never advances a clock, so it had never called repeat_due() at all. The row tests the latch, which the repeat does not touch. The prediction was reasoned rather than run.

What it hid is the real problem: the rate was about to ship into a harness with no coverage of this feature whatsoever, and that green line would have been read as coverage of it.

The fix was not to change that row. verify-input gains a repeat subject that holds a direction through the same latch and ticks repeat_due(), asserting these two numbers — a shape-only check ("it repeats eventually") would have passed on the 0.40 / 0.20 guess this port refused to ship. Five rows: nothing before the delay, the first repeat on the delay, the steady interval, cadence independent of frame rate (the code claims this in a comment, so it is asserted), and a direction change restarting the delay.

⚠️ The origin is one frame, and it is not a tolerance. repeat_due()'s first call only latches the direction; the clock accumulates from the call after it. In the port that first call is the frame the press is handled — the frame that produced the press-triggered step — which is the origin the finding measures its 12 frames from. Measured from t = 0 the harness read 0.433 against 0.402 and the tolerance would have had to be widened to hide a units mismatch.

🔴 It did not work on a real controller, and every instrument here said it did

Reported 2026-09-13, by a human holding a real stick: one step, then nothing.

held_direction()'s own comment said "Polled at the DEVICE, never through Input.is_action_pressed". It was not. For the stick it read _latched — a reconstruction of the stick's position from the event history — and only the d-pad and keyboard branches actually polled anything.

_latched changes only when an event arrives. A stick held perfectly still sends nothing, so the reconstruction is only as good as the last event seen, and any single event reading below RELEASE — a spring settling, a deadzone-shaped value, a driver emitting a zero on focus change — clears it with nothing afterwards to set it back. From then on the port believes the stick is centred while the player is holding it. That is exactly "it moves one item and stops".

Why nothing here caught it

Every instrument in this repo supplies the input it then measures:

instrument	what it feeds
verify-input's repeat rows	calls repeat_due() directly, after setting the latch through accepts()
--script=up,down,accept	InputEventAction — bypasses the input map entirely
--script=hold:down:2.0	injects one InputEventJoypadMotion

All three agreed with each other and none of them agreed with the controller, because none of them read a device. This is the same shape as the 2026-09-01 report that opened gamepad.gd: a synthetic-input test asserts the code after the input map, never the input map itself — one level deeper, and it caught us again with the lesson already written down.

What changed

held_direction() now polls Input.get_joy_axis() against the game's own 0.61, which is what its comment always meant. The latch survives as a fallback for injected events, because Input.parse_input_event() does move get_joy_axis() but the harness must keep working if that ever changes.

And --input-probe prints what the devices report, on change:

probe: [0] Generic X-Box pad Y=+1.000 X=+0.000 dpad=--  latched=1  held_direction=1  (ENTER=0.61 RELEASE=0.40)

⚠️ This fix is not verified. It is a defect that matches the symptom exactly, found by reading, and the only instrument that can confirm it is a human holding a stick. The probe exists so that the answer is a measurement either way.

What this does not claim

• That the repeat feels right. It runs now, and only a human can answer that: 0.134 s is ~7.5 steps a second, and the play-test that opened this asked for "slow enough to see which item is selected". If it reads as too fast, the 100 ms constant is Xenia's and not the game's, and no capture in that container would have revealed it.
• Any rate, or any bound on one. "Medium pace" is a direction, not a number, and it is not recorded anywhere as data.
• That the d-pad and the stick repeat at the same rate. Both repeat; nobody has said they match, and the code currently assumes one rate for both.