Fix check_for_low double-fire on button release

[Copilot]

Description

check_for_low() double-fires: when a held button passes the debounce threshold, it returns True but leaves _low_since_ms[key] populated. On subsequent release (pin high), the latch logic sees the stale timestamp and reports a second press for the same physical interaction. This causes the "create seed from camera" workflow to capture an image immediately on entry.

Root cause: The debounced-low return path didn't consume _low_since_ms[key], so the release-latch path (added to catch brief presses between slow polling calls) would re-fire.

Fix: Clear _low_since_ms[key] before returning True from the debounced-low path:

# Before (double-fires)
self.update_last_input_time()
return True

# After (consumes the press)
self._low_since_ms[key] = None
self.update_last_input_time()
return True

No screenshot — no UI changes; this is GPIO button debounce logic only.

This pull request is categorized as a:

• New feature
• Bug fix
• Code refactor
• Documentation
• Other

Checklist

• I've run pytest and made sure all unit tests pass before sumbitting the PR

If you modified or added functionality/workflow, did you add new unit tests?

• No, I'm a fool
• Yes
• N/A

I have tested this PR on the following platforms/os:

• Raspberry Pi OS Manual Build
• SeedSigner OS on a Pi0/Pi0W board
• Other

---

📱 Kick off Copilot coding agent tasks wherever you are with GitHub Mobile, available on iOS and Android.

[Copilot]

Pull request overview

Fixes a GPIO button-state leakage bug that could trigger an immediate “phantom” capture when transitioning into the camera workflow, and expands display support/handling (ST7735 + display switching) with supporting docs and tests.

Changes:

• Consume _low_since_ms[key] in HardwareButtons.wait_for() and add latching logic to check_for_low() to detect short presses between polling intervals; add unit tests covering held/released/noise/phantom scenarios.
• Add a new ST7735 (128×128) display driver, wire it into display configuration/settings and the display-driver factory, and add driver-focused tests.
• Add a home-screen very-long-press shortcut to switch display configs, plus renderer/keyboard/QR adjustments for dynamic canvas sizing and updated documentation.

Reviewed changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/test_io_config_profiles.py	Adds test coverage for ST7735 driver construction, SPI mode/flags, lazy init, timing, and EINVAL fallback.
tests/test_buttons_gpio_resolution.py	Adds latching + phantom-press regression tests for check_for_low() / wait_for().
src/seedsigner/views/view.py	Handles the new return code to re-render Main Menu after display switching.
src/seedsigner/models/settings_definition.py	Adds st7735_128x128 and desktop_128x128 display configuration constants/options.
src/seedsigner/hardware/displays/display_driver.py	Adds st7735 display type support and enforces 128×128 resolution.
src/seedsigner/hardware/displays/ST7735.py	Introduces the ST7735S 128×128 SPI display driver implementation.
src/seedsigner/hardware/buttons.py	Clears _low_since_ms on wait_for() returns and adds latching behavior to check_for_low().
src/seedsigner/gui/screens/screen.py	Adds RET_CODE__DISPLAY_TOGGLE, updates QR rendering sizing, and implements Main Menu long-press display switching.
src/seedsigner/gui/renderer.py	Adds ST7735 canvas-vs-display resize path and resizes outgoing frames when needed.
src/seedsigner/gui/keyboard.py	Makes default keyboard rect derive from the active renderer canvas dimensions.
docs/io_config.md	Documents shared pinout and the config value for the 1.44" ST7735 HAT.
docs/hardware_platform_support.md	Documents supported displays and the very-long-press display switching shortcut.

[Copilot]

QR images are now generated using (canvas_width, canvas_height). On non-square canvases (e.g. ILI9341 where canvas is 240x320), this will stretch the QR code via qrencode/PIL resize, which can reduce scan reliability. Consider generating a square QR image using a single size (e.g. min(canvas_width, canvas_height)) and centering/padding it instead of resizing to a rectangle.

[Copilot]

Same as above: next_part_image() is now called with (canvas_width, canvas_height), which will produce a non-square, stretched QR on portrait canvases. Use a square size (e.g. min dimension) and then composite onto the canvas to preserve module aspect ratio.

Suggested change

	image = self.qr_encoder.next_part_image(self.renderer.canvas_width, self.renderer.canvas_height, border=2, background_color=hex_color)
	canvas_width = self.renderer.canvas_width
	canvas_height = self.renderer.canvas_height
	qr_side = min(canvas_width, canvas_height)
	qr_image = self.qr_encoder.next_part_image(
	qr_side,
	qr_side,
	border=2,
	background_color=hex_color,
	)
	# Center the square QR image on the full canvas to preserve aspect ratio
	if qr_side == canvas_width and qr_side == canvas_height:
	image = qr_image
	else:
	# Use the same background color as the QR image
	background = "#" + hex_color
	image = Image.new(qr_image.mode, (canvas_width, canvas_height), background)
	offset_x = (canvas_width - qr_side) // 2
	offset_y = (canvas_height - qr_side) // 2
	image.paste(qr_image, (offset_x, offset_y))

[Copilot]

The PR title/description focus on the phantom camera button press fix, but this PR also adds a home-screen very-long-press shortcut for switching display drivers (and related display driver/settings/docs changes). Please update the PR description/title to reflect this added scope, or split into separate PRs to keep review risk isolated.