GENERATED VIEW · PRIVATE ARCHIVE · DO NOT SERVE
RobCo Industries · Archive Museum

primary-users-devices.md



MEMORY

memory/localmode-1060dea9-0dc3-42b4-958d-2a6f7a6c0c58/primary-users-devices.md

sha256 27f015e29e57e239 · 2319 bytes · original held in the private archive

--- name: primary-users-devices description: "The two primary RobCo users + their exact devices. Owner = OnePlus 15. Brother = OnePlus 12R and is MAXIMUM PRIORITY. Both Android/Chromium — which materially shifts the storage-eviction risk picture." metadata: node_type: memory type: user originSessionId: 0b7c5f41-bb34-4ab4-93be-fdb363f5eb40 --- **Told 2026-07-14.** The two primary users and their devices: - **Owner (zerckzzy) → OnePlus 15.** - **⭐ Brother → OnePlus 12R, and is MAXIMUM PRIORITY.** When a device tradeoff has to be made, the 12R wins. Both are **Android / Chromium** phones. This is durable targeting context, not trivia. ## ⭐ WHY IT MATTERS — it shifts the storage-eviction picture (honestly) The `navigator.storage.persist()` work was justified largely by **an iOS-Safari eviction warning** — iOS aggressively evicts localStorage/IndexedDB. **Neither primary user is on iOS.** Android Chromium is *far* less eviction-happy and grants persistence readily (especially for an installed PWA). So: - The persist() fix is still **correct and universal** — keep it; future/other users may be on iOS, and it costs nothing. - But the *scary* case (silent hot-save eviction) does **not** materially threaten either primary user today. Don't over-weight iOS-specific behaviour in future decisions; the real test surface is **Android Chrome / installed PWA**. ## ⇒ THE SUPPORTED BROWSER/PWA BOUNDARY The declared support boundary has a concrete primary target: **Android Chromium, OnePlus-class hardware, PWA-installed.** The 12R is the device to verify against first. ## SCREEN SIZE OnePlus 12R/15 sit at roughly 412+ CSS px wide, which the existing mobile-baseline breakpoints already bracket — no device-specific breakpoint is needed for them. But auth/PWA-specific behaviour must be verified on a **real** OnePlus, not just at an emulated width. (Current breakpoints and render-check widths: read them from the gate config, not from here.) *Corrected 2026-07-20: removed the commit ref, the hardcoded render-check widths, protocol numbers, and a version-tagged audit reference. Device facts and the eviction reasoning are unchanged.* Related: [[audit-completeness-gaps-2026-07-14]] (the persist finding), [[robco-uos-project]], [[design-overhaul-direction]] (mobile-primary).
STAMP · generated for RELEASE v2.8.5 commit 06e5180 (06e51801b38a) · archive input-tree hash c07fbfbdd2e1ddeb · 754 files · no wall-clock timestamp (regenerates identically when nothing changed).