The script's set -euo pipefail made it die without any visible error in
two places: an empty /tmp/jtube-adbkeys/ failed the glob cat (the error
itself was discarded, so the output just stopped mid-script — this hits
any host after a reboot), and the boot-wait loop was killed on its first
iteration because adb exits non-zero while the device is still booting.
Both were swallowing the whole && chain (adb connect, scrcpy), making it
look like the emulator was broken when it was fine or about to be.
The stale bridge-emulator.sh duplicate (different port, broken vendor
keys path) is removed in the same spirit — keeping it invites running
the wrong script. Docs updated to match the one script.
Document how the host emulator is driven over adb from the agent
container: the start script now boots headless by default and trusts
remote adb keys via ADB_VENDOR_KEYS at launch, since the TV system
image is a user build that blocks pushing keys to adb_keys after boot.
AGENTS.md points to the new doc and fixes the stale SDK path.
Perf work for the TV FPS complaint: every page that loads from the API
also warms Coil's disk/memory caches (video thumbnails at the same
480x270 the card renders, channel banners/avatars) so rows scroll
without popping in over the network. Prefetch keys match render keys
exactly so cache hits are guaranteed.
bridge-emulator.sh runs OUTSIDE the container on the emulator host:
boots jtubesc if needed (ADB_VENDOR_KEYS whitelists the host adb key),
installs/starts JTube TV and binds :5557 -> adbd 5601 so the container
and LAN can drive it. Run it once in your rutherford shell.
On launch and on every destination switch the first card of the
content gets focus automatically - invisible focus on chrome confused
'Do I need to press DOWN first?'. The search button shows a white ring
when focused, and the top bar gained a real account button (person
icon + logged-in username) replacing the cryptic monitor icon.
start-emulator.sh: headless-safe boot (no -gpuhost crash without a
display), whitelists adb public keys headlessly (drop them into
/tmp/jtube-adbkeys/), and starts the socat bridge itself.
scrcpy cannot talk to headless emulators inside the container: the
AOSP ATD image lacks framework pieces (KeyCharacterMap NPE) and there
is no GPU/KVM in the container at all. The script boots the android-tv
image image on the host with -gpu host and a window (so the RSA dialog
is clickable), installs the current debug APK and leaves the app
running; the container's socat bridge then exposes it as
rutherford:5556 for adb/scrcpy from anywhere.
Paused playback now shows a bottom panel with channel + title + views
+ upload date and position instead of the controller bar (which hid
itself while paused - it was colliding with custom info). The
controller no longer flashes 4 s at playback start: it only reveals
on UP/DOWN-style keys.
For the TV frame-rate complaint: thumbnails decode at 480x270 RGB_565
via a global Coil loader and avatars at 160px, instead of full-size
1080p+ jpegs re-composited per scroll frame. Measured on the
software-rendered emulator: stable 10% janky / p50 18ms during row
scrolling; real-hardware measurement via scripts/fps-check.sh when
the couch is free.
BACK from a Home row also refocuses the watched card now (row-local
focus requesters + restore token), fixing the 'focus lands two-three
cards left' complaint.
Releases need more than eyeballs: scripts/e2e-api-check.sh now runs
18 assertions over the live instance (login/refresh, listings, private
rows, subscriptions incl. remote channels, watching + stored position,
file token, playlist reachability) and TESTS.md defines the manual
D-pad checkout that a script cannot do.
Also switch the channels screen to the user's own channels (the
search-based local listing rejects empty queries) and resolve the real
account name at login since logins accept email addresses.
