# Attempt to reach a .tbm-bearing submenu -- NOT REACHED, and why.
#
# 2026-08-30. The open question is ui-forced-backdrop.md's 24 .tbm deciders:
# our compose draws NOTHING for a .tbm, so their verdicts are 'correct or
# inert' and indistinguishable. Every main-menu destination except EXTRAS
# lands on an archive holding one, so no focus detector is needed -- press A
# on whatever is focused and identify the screen from the capture.
#
# RUN 1 -- the title->menu transition was TIMED (tap, wait 8 s, assume).
#   8 s later the screen was still the TITLE: glyph 714, the plate pulse's
#   trough. So the second tap performed the transition and the 'submenu'
#   capture is the main menu. Void for the question asked.
#
# RUN 2 -- menu DETECTED instead of timed (glyph in 250..420 for 6
#   consecutive samples; the main menu is 327 and the plate is 714..1520).
#   [351.3s] tapped A
#   [357.2s] MENU detected (glyph 327) -- pressing A into a submenu
#   [358.0s] tapped A
#   [369.8s] submenu captured (glyph 327, mean 44.4)
#   -> still the main menu. 2.4 % of pixels differ from the menu capture.
#
# 🔴 WHY: the second tap was NEVER DELIVERED.
#   [file-pad] vk=5800 lines: 2  -- that is ONE press (down + up)
#   [RE-INPUT] -> user=0 vk=5800: one down, one up
#   swallowed by IsUIActive: 0   (so NOT the sign-in dialog path)
#
# The tap at 358.0 s came 0.8 s after the menu appeared, while the guest was
# still loading it. A 0.12 s press is missed entirely if the guest does not
# poll during that window -- the pad driver reports what it emitted, and the
# guest simply never asked.
#
# WHAT WORKS AND IS NOW VALIDATED:
#   * the plate-pulse title detector (three runs)
#   * the MENU detector: glyph 327, matching live-main-menu.png exactly
#   * delivery is checkable in-log: [RE-INPUT] ... -> user=0 vk=5800