K2GO-380 docs(post-rebrand): device verification report + global test matrix - #534
Merged
luisguzman-adfa merged 11 commits intoSep 3, 2026
Merged
Conversation
…l test matrix (WIP) Records the org.iiab.controller -> org.appdevforall.k2go rebrand device sweep: static rebrand-safety recon, the global test matrix (identity- sensitive first, then functional), a live-results log, and finding F1 (uninstalled content cards read RED instead of GRAY after a backup — pre-existing status-derivation bug the rebrand's clean install exposes). Live results so far all PASS: FileProvider registered under the new package, content endpoints serving, OTA host reachable, Firebase inits under the new package. WIP; UI/heavy checks pending.
…spot QR) PASS on device Live UI checks on the OnePlus 7T (release build, org.appdevforall.k2go): - I2 FileProvider runtime: Clone->Send->"share another way" opens the Android share sheet (base.apk); getUriForFile resolves under the new authority. ADR-5368 check 10 independently confirmed. - F-m: hotspot join QR renders, LocalOnlyHotspot starts, get-app QR renders under the new package. K2GO-375 fix intact.
…d on base tier Device run (OnePlus 7T, debug build, base tier): clean install + base system install, full backup (1.65 GB, k2go_ prefix, DeepOpService), Finish -> Home all PASS under the new package. F1 (uninstalled cards RED after backup) did NOT reproduce on base tier -- cards read GRAY correctly; the red window is a narrow post-restart transient, likely wider on slower/full -tier restarts. Retracted an earlier backup-naming false alarm (mis-tap on an old file, not a real inconsistency).
…orking conclusion Full-tier retry (2.43 GB backup, screen-recorded Finish->Home): all cards Ready immediately, no red flash. F1 (red-instead-of-gray after backup) did not visibly reproduce on device across base or full tiers on this build; Home renders after services answer, so cards resolve straight to green/gray. The code defect stands by inspection; the visible symptom needs the reporter's exact build/tier/timing to reproduce.
…e checks pass; pristine verdict Device batch (debug, full install): private dir under new pkg (3), cold boot UP with 0 kills via K2Go-Reconciler (8), proot runs (9), FG watchdog notification under new pkg+channel (14), maps landing + FqrController overlays render (18), legacy intent actions drive live flows (13). Every single-device ADR §10 check passed; only OTA new->new (11) and 2-device clone (15) remain, needing external setup (both reasoned-not-observed in the ADR). Restore (16) and dashboard rebuild (17) optional. Verdict: the rebranded APK behaves as pristine for single-device operation.
…only OTA new->new remains Monitored a real 2-device clone on the OnePlus (receiver): pairing with the peer, CloneShareService FG on clone_channel, the CLONE holder quiesced the server (reconciler desired=DOWN holder=CLONE), rsync transfer, and the received system booted healthy (home=301, kiwix=200) with the holder released back to UP [holder=NONE], 0 kills. Benign non-rebrand SELinux denial noted (librsync ioctl TCGETS on a pipe). Every single-device ADR §10 check plus clone now confirmed; only OTA new->new (11) remains, needing an update server. Verdict: the rebranded APK is pristine.
…2 progress-reset bug Monitored a live dashboard rebuild: reconciler held holder=DASHBOARD ~2.5 min (server never dropped, 0 kills), released to holder=NONE on completion -- DashboardRebuildService sound under the new package. Surfaced F2 (not rebrand-related, K2GO-95 area): the rebuild progress bar resets to 0 on return from the notification and the notification shows no percentage, because progress is computed fragment-locally from a wall-clock that resets on recreation and the service broadcasts only coarse state. Design read: progress has no persistent owner.
…nfirmed + filed as K2GO-381 Restore round-trip (check 16) PASS under the new package (DEEPOP_RESTORE, RESTORE holder quiesce, TarExtractor 2.43 GB, healthy boot, 0 kills). Logged deep-op UX cluster F3-F5 (restore ETA, notification deep-link target, run-in-background nav) and F6: content services don't self-heal after a deep-op because service-heal.ts WATCHED lists only kiwix -- root cause confirmed in code, filed as K2GO-381. F6 also reframes F1 (the red/Unavailable tile is a down service, correct display, not a bug).
…verdict; OTA deferred to 0.9.0 Post-rebrand device verification complete: every single-device ADR-5368 §10 check plus a two-device clone passed on the OnePlus 7T; the rename changed identity and nothing else and the app runs correctly across every exercised mechanism. Only OTA new->new remains, deferred to a publishable 0.9.0. Findings F1-F6 recorded with owners (F6 -> K2GO-381; F2-F5 and a backup/restore progress+cancel standardization to be filed).
…n and passed, not "optional" The interim verdict in section 3.5 still said checks 16 (restore) and 17 (dashboard rebuild) "remain optional deeper round-trips" -- written before they were run. Both were subsequently exercised on device and passed (section 3.4 and section 4); correct the line so it no longer contradicts the rest of the report.
… minor fixes
Full read-through for internal consistency. Main fix: the F1 section
described F1 as a confirmed display defect ("the code defect is still real
... the fix stands"), which contradicted F6 and the conclusion that F1 is
reframed as a down service correctly rendered. Rewrote F1 as
investigated-and-reframed: the 404->gray / 502->red split (shared with
service-heal.classifyProbe) is correct, a not-installed platform reads
gray, and the reported red was a down service (F6). Also aligned the §3.3
F1-repro row and F-g note, softened the I12 "release gate open" wording to
match its PASS result, dropped "API level TBD", and changed §4 "runnable"
to "exercised" noting check 12 is covered by inspection.
luisguzman-adfa
deleted the
docs/K2GO-380-post-rebrand-device-verification
branch
September 4, 2026 13:01
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Adds
controller/docs/POST_REBRAND_DEVICE_TESTS.md— the on-device verification of theapp-identifier rebrand (
org.iiab.controller→org.appdevforall.k2go, K2GO-293 / ADR-5368):a static rebrand-safety recon, a global device test matrix mapped to ADR-5368 §10, a
live-results log, and a findings log. Doc only, no code changes.
Outcome
Verified end to end on a OnePlus 7T (base + full tiers, plus a real two-device clone). Every
single-device ADR-5368 §10 check passed — fresh install, boot (0 spurious kills), FileProvider,
notifications, custom-View screens, intent actions, content serving, backup, restore, clone,
dashboard rebuild — so the rename changed identity and nothing else. Only OTA new→new is
deferred, pending a newer published build (0.9.0).
Findings (none caused by the rebrand)
service-heal.tswatches onlykiwix) → filed as K2GO-381.