examples: run #9's test plan on hardware

[bdbarnett]

PR #9 (merged as edea99c, 2026-09-07) added nine examples and rewrote uvc_display.py to prefer MJPEG through jpegio. It was merged on a static review: every _usbif call checked against the C module's locals, argument tables and tuple layouts, the usbif.uac_audio / usbif.hid_keyboard / events surfaces checked, all twelve files compiled, and host_enum.py run on desktop CPython (where it reports no capabilities on WSL, as expected). Nothing was run on a board.

The PR's own test plan, still unchecked:

• P4: soundcard.py — Speakers appear on a PC; audio audible (2026-09-23 on 1.29/IDF 5.5.4: Brad, "speaker works perfectly")
• S3: usb_speaker.py hosting the P4 — tone on the P4 speaker (2026-09-23 after device: size every bulk endpoint for the negotiated speed, not only MIDI's (usbif#24) #27, host: USB Audio 2.0 devices, a FIFO split per session, full-speed audio packets (usbif#28, usbif#2) #30: 48 kHz stereo through the P4's 2.0 sound card; Brad heard the two-second tone)
• hid_keyboard.py / hid_mouse.py against a PC -- P4 both (2026-09-23, typed and moved); S3 the mouse example after device: number endpoints per costume; refuse a costume past the controller's IN budget (usbif#23) #25 (cursor out and back, no refused report); the keyboard example shares the endpoint and was not rerun on the S3
• hid_host.py with a keyboard (M1 events)
• uvc_display.py with jpegio firmware (MJPEG) and without (YUY2) — this one was a working example before the rewrite, so it is the first to run
• host_enum.py on a board
• usb_serial.py, usb_mic.py, ram_drive.py
• costume_selftest.py still 31/31 (2026-09-23: six functions now, so 63 costumes; P4 63 checked 0 failed; S3 46 wearable + 17 refused for the S3's four IN endpoints, by design after device: number endpoints per costume; refuse a costume past the controller's IN budget (usbif#23) #25)

Two things a run may surface that the static pass could not: soundcard.py reads I2S pins and codec hooks from board_peripherals under several guessed names and falls back to _output_power / _codec_call; and usb_mic.py / usb_speaker.py read the negotiated format through getattr(..., "format") with a .stream fallback. Whichever branch a real board takes should be the one that stays.

[bdbarnett]

Running the plan on the 1.29 images (2026-09-23), boards on the PC:

• hid_keyboard.py / hid_mouse.py on the P4: pass. typed 'hello from usbif': ok, hid_leds: 0x01; the mouse moved +x and returned, done; cursor should be back where it started.
• The same two on the S3 Touch: blocked by S3 device HID: exactly one report goes out after a costume change, then the IN endpoint is never ready again (P4 unaffected) #23. After a costume change exactly one HID report goes out on the S3 and the IN endpoint is never ready again, so the key-down goes and the key-up is refused. The P4 is unaffected. Filed with the twenty-report trace.
• soundcard.py on the P4: the costume is CDC+UAC (303a:4011) and Windows lists an Espressif Device under MEDIA. The audible half is being checked now.

The S3-hosting-the-P4 line (usb_speaker.py) follows once the sound card is heard.

[bdbarnett]

soundcard.py on the P4: pass. Brad selected the Espressif Device as Windows' output and played through it: "speaker works perfectly". Two of the plan's four lines pass on the P4 (this and HID); the S3's HID lines wait on #23; the S3-hosting-the-P4 tone is next.

[bdbarnett]

usb_speaker.py on the S3 hosting the P4: blocked by #24. The S3's host sees the P4 behind the hub and the P4 stalls endpoint zero at full speed in every costume, so no audio output is found. Where the plan stands on 1.29: P4 sound card pass, P4 keyboard and mouse pass, S3 keyboard and mouse blocked by #23, S3 hosting the P4 blocked by #24.

[bdbarnett]

S3 HID line unblocked by #25 (endpoints numbered per costume; the S3 drives IN endpoints 1..4). hid_mouse.py on the S3 Touch: costume worn, +x and -x moves, no refused report. Still blocked: "S3 hosting the P4" and usb_speaker.py (#24, the P4 as a full-speed device).

[bdbarnett]

Ticked today's lines: P4 sound card on a PC (heard this morning), the S3 mouse example, the S3 hosting the P4 sound card (#27 + #30, tone heard), and the costume self-test (63 on the P4, 46 wearable on the S3). Not run today: hid_host.py with a keyboard, uvc_display.py, host_enum.py, usb_serial.py / usb_mic.py / ram_drive.py.