Files
Kaj SchittecatandClaude Opus 5.5 9d0301b80a touch: beta_92, faster screens on most boards, Arabic fixes, GitHub fixes
Frame rate: full-width PSRAM draw band (6 bands a frame instead of 14 on the
T-Deck), the still aurora drawn once and copied, LVGL at -O2 with its blend and
label loops in IRAM (scripts/build/perf_lvgl.py, 1 KB more loop stack), circle
and gradient caches, newlib memcpy. Rolled out to the M9, Attaky, RAK, Wio,
CrowPanel and V4-R8; the Pager gets the band only, the V4 the caches only.
Frame benchmark tour (DOC_PERF_BENCH) for measuring it.

Fixes: hop counts read from the packed path length (#621); Arabic shaping
rows patched in LVGL (patch_lvgl_arabic.py), Arabic 101 shift layer, kaf,
symbols, text sent as standard letters (#582); info popup colour runs that
never wrap (#602); keypad nav reaches the chat header (#623); the browser
opens the changelog (#622, gen-changelog.py at every site deploy); the
Theme background colour picker no longer edits the accent; flasher entries
for P4 units with factory ESP-AT Wi-Fi chips (#612, #619).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-09 11:53:33 +02:00

21 lines
2.8 KiB
Plaintext

# beta_92 - faster screens on most boards, Arabic fixed up, and a round of GitHub fixes. TEST BUILD.
# One user-facing note per non-blank, non-# line; # lines are section comments.
# Order: what this build is, speed, then the fixes.
Test build, on top of beta_91. The stable build stays beta_85.
# --- Faster screens ---
The screens draw faster. On a T-Deck with the still aurora, Home redraws in 85 ms instead of 126, the app drawer scrolls at 69 ms a frame instead of 102 and a chat list at 49 instead of 70: a third to a half more frames a second, with the same pixels.
How: the T-Deck drew every frame in 14 slices, each a walk through the whole screen, and now draws it in 6; the still aurora is drawn once and copied instead of worked out again every frame; and the drawing code is compiled for speed, with its busiest loops in the chip's fast memory.
The same treatment reaches the ThinkNode M9, the Attaky, the RAK WisMesh Tap V2, the Wio Tracker L2, the CrowPanel 3.5 and the Heltec V4-R8, and the T-LoRa Pager draws in fewer slices. Those were not tested on their own hardware yet: please report how scrolling feels there, and anything that looks off.
The Heltec V4 gets the smaller part (rounded corners and gradients are cached now); it has no room left for the rest.
# --- Fixes ---
Hop counts above 64 are gone (#621, reported by jesshampshire). With 2- and 3-byte path hashes the hop count shares its byte with the hash size, and a few places read the whole byte: the Discovered list, the "auto-delete above N hops" setting (it could throw away nodes one hop away) and the checks for direct neighbours.
Arabic, thanks to jcyrio and a native-speaking tester (#582): letters join properly now (آ, ئ, ة and the Persian digits were wrong in the graphics library), Shift gives the Arabic 101 shift layer (أ إ آ, لأ لإ لآ, ، ؛ and the tashkeel), ك is on 8, ! $ @ type themselves again and ? types ؟. Text leaves the device as standard Arabic letters, so a #channel name gives the same channel as on a phone.
The message info popup keeps its colours when a long repeater name wraps (#602). NewtsLament found the cause.
On the keypad boards (ThinkNode M9, Tanmatsu, T-LoRa Pager) a chat's settings, with Region & scope, can be reached again: move up to the conversation header and press Enter (#623, reported by bobbythefett).
The web browser opens the wadamesh changelog when there is no home page on the SD card and Wi-Fi is up (#622, an idea from jesshampshire). It is also on the site: wadamesh.com/changelog.html.
Settings > Theme > Background > Colour changed the accent colour instead; it sets the background now.
T-Display P4 with Wi-Fi that never connects: the flasher now offers "factory Wi-Fi chip" images, for units whose Wi-Fi chip still runs LilyGo's original firmware (#612; jalevdev spotted the missing choice in #619).