Files
Gathin23 7d6178537d tools/gen_tables.py: the TV tables, brand line and counts come from examples/*/tv.json
Every list PR added a row to the same two tables (README.md and examples/README.md), so each merge
made the other open list PRs conflict. Each folder now carries its row in tv.json (tv, os, brand,
notes), and the script rewrites the tables between <!-- tv-table:start/end --> markers, the brand
line between <!-- tv-brands:start/end -->, the card's alt-text counts, the languages row and STATS
in tools/readme_stats.py. Everything outside the markers stays hand-written.

The 260 tv.json files are backfilled from main's tables, so both tables come out byte for byte as
they were: "readme_after"/"examples_after" pin the 31 and 42 rows that sat out of alphabetical
order, and "readme_os" keeps the two README OS cells that differed from examples/README.md. The
brand line catches up with six lists it had not counted yet (Hisense, Sony and Xiaomi one more
each; INCLO, KTC and Liberton); the alt text already said 260 TV models and 77 brands.

--adopt PATCH turns the rows an older PR added to the tables into its tv.json, so open PRs can
merge without their README hunks. On a PR the tests check each tv.json; on main
(CLEANYOURTV_STRICT) the committed tables must be what the script writes. The package-fate and
device-note checks read the notes from tv.json, so a list is checked before its row exists.
2026-10-24 15:43:20 +05:30

7 lines
515 B
JSON

{
"tv": "Philips 55OLED708/12 (TPV, PH1M_WW_9972, MediaTek), Poland",
"os": "Google TV 12",
"brand": "Philips",
"notes": "Google TV generation, not the 707's board: DAC analytics, frameworklogger, Samba TV ACR, NFT Collection, Play-Fi, Control4, Matter, Mimi, Alexa, Chromecast, setupwraith (Wi-Fi re-joined first) and the Philips launcher tab; `demome` holds TV_INPUT_HARDWARE so it stays; the remote's gear and Apps buttons need HomeKeyService (gear by scancode 141); nothing re-enabled after a reboot"
}