Skip to content

CONTRIBUTING, install-phone: phones without A/B slots, and never flash a bootloader to match slots - #383

Merged
derekross merged 1 commit into
mainfrom
contrib-non-ab-installer
Nov 3, 2026
Merged

derekross merged 1 commit into
mainfrom
contrib-non-ab-installer

Conversation

@derekross

Copy link
Copy Markdown
Owner

Two things, both from tonight's reviews:

  1. The installer told people to flash the bootloader. tools/install-phone refused differing bootloader slots with "flash the same factory bootloader to both from stock first", against CONTRIBUTING §3.1. Issue cheetah (Pixel 7 Pro, SKU GE2AE): touch never identifies, and the panel loses TE after wake with screen_off=2 (both fixed locally, details inside) #332's reporter did exactly that. The installer, device/cheetah/REINSTALL-RUNBOOK.md and §3.1 now say: never flash a bootloader partition; sideload the full OTA twice from stock, else ask in an issue.
  2. Phones without A/B slots get their own rules in §3.1. Keeping stock on slot A is a development-era safety net, not a requirement. A single-slot phone (such as the Mi 8 in device/dipper: record stage-0 facts and first light #330) can become a supported device under rules written here first, never inside the device's own PR: named boot partitions plus userdata only, through our tools; a full read-only backup and written-back steps before the first flash; calibration and identity partitions never mounted; and a reviewed answer for the initramfs's no-USB-host panic.

bash -n tools/install-phone and tools/pr-check origin/main pass.

🤖 Generated with Claude Code

https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d

…h a bootloader to match slots

The installer refused differing bootloader slots with "flash the same
factory bootloader to both from stock first", which contradicts §3.1's
"never flash bootloader". A reporter on issue #332 followed it literally
and flashed bootloader_b. The installer, REINSTALL-RUNBOOK and §3.1 now
give the sanctioned way: sideload the full OTA twice from stock so each
slot gets the same build, else stop and ask in an issue.

§3.1 also gains "Phones without A/B slots": keeping stock on slot A is a
development-era safety net, not a requirement for every phone. A
single-slot phone can be supported under its own rules: a device entry
written here before any flash (never self-granted in a device PR), only
the named boot partitions and userdata written through our tools, a full
read-only backup with the write-back steps read before the first flash,
calibration and identity partitions never mounted, and a reviewed answer
for the initramfs's no-USB-host panic, which would otherwise restart into
the same image.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d
@derekross
derekross merged commit 3cc6525 into main Nov 3, 2026
2 checks passed
@derekross
derekross deleted the contrib-non-ab-installer branch November 3, 2026 05:34
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…h a bootloader to match slots (derekross#383)

The installer refused differing bootloader slots with "flash the same
factory bootloader to both from stock first", which contradicts §3.1's
"never flash bootloader". A reporter on issue derekross#332 followed it literally
and flashed bootloader_b. The installer, REINSTALL-RUNBOOK and §3.1 now
give the sanctioned way: sideload the full OTA twice from stock so each
slot gets the same build, else stop and ask in an issue.

§3.1 also gains "Phones without A/B slots": keeping stock on slot A is a
development-era safety net, not a requirement for every phone. A
single-slot phone can be supported under its own rules: a device entry
written here before any flash (never self-granted in a device PR), only
the named boot partitions and userdata written through our tools, a full
read-only backup with the write-back steps read before the first flash,
calibration and identity partitions never mounted, and a reviewed answer
for the initramfs's no-USB-host panic, which would otherwise restart into
the same image.

Claude-Session: https://claude.ai/code/session_01BeEcdVeSfyLCbijF8XQF1d

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant