chore: v2.11.1

[angeloINTJ]

What

v2.11.1: the version, and the documents that name the current release. The maintainer chose the number on 2026-10-05: a patch release of fixes, #255 and #256.

SIMUT_VERSION 2.11.0 → 2.11.1 is the only source change. The two strings have the same length, but the login page carries the version, and its gzip came out 5 B smaller. Rebuilt from main de13562 plus this commit, the seven images measure:

• release, test, asserts and pico2_w_release: 8 B under the record, measured and .bin;
• alpha, air and test_https: equal to the record, because their alignment absorbs it.

nm shows the login page as the only symbol that changed. tools/flash_budget.json takes the new numbers, and §3.1 of PLANO_STABLE.md takes the four rows that changed. No budget moves.

The release notes

The v2.11.1 sections of CHANGELOG.md and CHANGELOG.pt-BR.md cover #255 and #256:

• The panel's accounts:
  • SAIR skipped the forced PIN change (finding 9);
  • the account editor granted bits the account at it did not hold (finding 8);
  • a stricter PIN minimum locked shorter PINs out of the panel;
  • the too-short line always said 4.
• Sensors: "Adopt the probe" read the wrong GPIO and left the sensor in quarantine (finding 1).
• The setup access point: the one the boot gesture opens lived by leftover memory (finding 25).
• Updates: the Files page now says why an image is refused with v=6.

Signed .bin against the published v2.11.0:

Image	Size	OTA headroom
release	1,032,933 → 1,033,357 B	7,027
alpha	990,621 B, unchanged	49,763
Air	1,031,485 B, unchanged	8,899
Pico 2 W	997,001 → 997,417 B, in a slot of 1,568,768 B	—

Known keeps v2.11.0's open items, minus two:

• the Files page's silent v=6, which fix: four high findings of the manual review, and the reason for v=6 #255 fixed;
• a try (_nosave=1) of a stricter PIN policy that does not mark the PINs below it. It still marks no one, but since fix(panel): a PIN below a stricter policy still identifies, and the too-short line says the minimum #256 a PIN shorter than the policy goes to the change without the mark, so the item has no effect left. The fix's entry says so.

It adds one: a narrower PIN alphabet, or a lower length ceiling, still strands an old PIN (#256).

Version mentions

• The three READMEs: the current release and the manual row.
• docs/MANUAL.md and its pt-BR twin (the header), and docs/README.md.
• The landing page, in en, pt and es: the v2.11.1 card first and the v2.9.0 card out.
• The product manual, regenerated:
  • chapter 1: what v2.11.1 changed;
  • chapter 3: the release's file names;
  • chapters 4, 9, 13, 14 and 28, and appendix C.

Two rows of chapter 28 were behind the code. They now say:

• the access point opened by the boot gesture follows the 15-minute limit like the others;
• identification at the panel takes any PIN from 4 characters.

Checked locally

• every step of the CI gates job (38, the PlatformIO install aside), including the nine native suites (639 cases), plainly and under ASan and UBSan;
• check_flash_budget.py on each of the seven images, and with --table;
• build_manual.py --check;
• scan_secrets.sh.

Next, before the tag

The release gate runs on the candidate the CI signs from main, after this merges and the maintainer approves the signing job:

1. The Pico W, on the bench now: the gate of AGENTS.md §2, steps 1 to 4.
2. The Pico 2 W, which the maintainer puts on the bench when asked:
   • the candidate over the air, from the bench image: it boots on trial, keeps itself, and keeps the configuration;
   • then the factory .uf2, dragged onto its BOOTSEL drive by the maintainer. The board is an RP2350 A2, and this is the erratum E10 case.

The results go in a second PR, as in v2.11.0 (#253), and its merge is the tag.

🤖 Generated with Claude Code

https://claude.ai/code/session_01BniVDEDSHGQgcoqmCRn5um