Y2Linux 1.0 Beta 8.14 (Baseline 111)

Status: PUBLISHED as GitHub release baseline-111 (2026-10-20). Release 1.0.0-beta.8.14, Reborn 1.0.0-beta.8.14, kernel 6.18.56-y2linux-1.0.0-beta.8.14, update sequence 7800, Linux 8e21366, Reborn 5cda8bb. Built with tools/production/release.py (130 s, resumed workspace, update based on baseline-110): container suite 842 OK, Reborn tests and Clippy (-D warnings) clean; all seal checks PASS, no MediaTek firmware in any image. Build 111 itself was installed on the maintainer's player over the air and checked by the owner (below). Owner: "build 8.14 with 1–4 and publish it". Milestone audit BETA-8.14 in roadmap-gap-audit.

Image	SHA-256
BOOTIMG	`71db0088e3f5fb62fa32699a70636a754834c8c282dfe13fe2ab514cde21bbcf`
Y2ROOT	`25301d97268bf24d0dd3dc22cb3d9cc52d92849ca729df683e5653813058de96`
RECOVERY	`445bee09645e4d2a4beac71bc0f178e85366af1a770b291d09660a5cc0c9b26e`
Y2DATA (no SSH key)	`ce7d1b76dbc06d1319aedf0b8e9b37d746e0dd5100a3e5910c0f797bc2266d3c`

What changed since Beta 8.13

Request / report	Cause	Change
"Every shuffle is the same order"	The shuffle seed was the start song's ID plus its index (`main.rs` `play_members`), and Shuffle rows always started on the first song: identical order every time	Seed from `getrandom` per shuffle (`reborn_platform::random_u64`), splitmix64 with a 64-bit range reduction; a Shuffle row lets any song start; a song picked with Shuffle on still plays first. Reborn `2861ec9`, tests `shuffle_and_seek_tests`
"Seeking jumps ~30 s per tick"	A held Previous/Next sent `Seek(±30 000)` every 90 ms (≈ 5.5 min of song per second)	The UI sends the direction; the step is 1 s for the first 1.5 s held, then 3 s, then 10 s, at most 1/60 of the song (Reborn `2861ec9`)
"All Songs and Shuffle on Genres"	All Songs existed since 8.8; Shuffle did not	A Shuffle row under All Songs on a genre (also in Cover Flow)
"Long titles push into the menu"	The Home card's now-playing title was drawn unshortened at x=356	Fitted to 200 px with an ellipsis, like the line below
Report bf017f02: `storage_error` "I/O error, dev mmcblk0 … WRITE_ZEROES" during the Internal Storage setup	The eMMC advertised TRIM, so mke2fs's zero-out became an erase, which the MSDC guard refuses (`storage-policy.h`); the kernel then wrote zeros, data fine	Overlay 0009: the guarded eMMC marks TRIM broken in `init_card`, so zero-outs write zero pages (Linux `f9a3bd8`)

Device results (build 111, maintainer's player)

Check	Result
OTA 8.13 → 8.14 (stage + apply)	Acknowledged 7800, Reborn health ok
`/sys/block/mmcblk0/queue/write_zeroes_max_bytes`	51 904 512 on 8.13 → 0 (discard still advertised, 104 MB, unused: `nodiscard`, no discard mount, no fstrim)
Shuffle twice on an album, held Next, genre Shuffle row, long title on Home	checked by the owner on the player: all as intended

Known gaps

The seek schedule is a first tuning; it may change with more feedback.
Japanese/Chinese/Korean text still shows "�" (planned next: a Noto Sans CJK fallback with a runtime glyph cache).
USB DAC output needs a hardware check of the port's 5 V supply first; video playback is not planned for 1.0.
Unchanged from 8.13: Strawberry itself not run, Windows Explorer display of music Names, slow crystal, sleep with a
parked CPU, #47 confirmation, WAF rule and billing alert, sleep drain (#42), GC9503V confirmation (#45).