Master findings table
> **Historical record.** The dates, candidate identity, "current" claims, > next steps and permissions below belong to this recorded boundary. See > [current state](../../CURRENT_REBORN_STATE.md) for the latest physically observed result.Read scope/labels before interpreting a row. Classification is the decision category; Evidence state uses the requested VERIFIED/IMPLEMENTED/PARTIAL/WEAK/MISSING/UNKNOWN labels. Confidence concerns the stated finding, not unperformed hardware tests. HIGH severity does not mean a physical failure was observed. No newly demonstrated protected-partition destruction or battery incident is alleged.
P0 = before more feature work; P1 = before public testing; P2 = before stable; P3 = cleanup/future. “Conditional” means blocker only if that capability is offered. GOOD rows use LOW severity because they are preservation recommendations, not defects.
Correctness and release risks
| ID | Subsystem | Finding | Classification | Evidence state | Severity | Confidence | Concrete evidence | User impact | Recommendation | Release blocker? | Suggested order |
|---|---|---|---|---|---|---|---|---|---|---|---|
| F01 | Native media | Moved-from canonical AVFrame is not freed | IMPROVE | WEAK | HIGH | HIGH | media.c, source_to_canonical ~539–564; pinned FFmpeg buffersrc.c:av_frame_move_ref; host RSS probe | Long-session accumulation | Repair ownership and prove bounded live allocations | Yes | P0.1 |
| F02 | Playback/ALSA | Plan opens exclusive PCM before current worker stops; ordinary controls rebuild playback | IMPROVE | WEAK | HIGH | HIGH | main.rs, load, volume/seek effects; audio.c, rb_alsa_plan; playback.rs, load_with_gapless | Controls can fail or interrupt sound; source-derived EBUSY risk | Serialize planning/reconfiguration under sink ownership and test rollback | Yes | P0.2 |
| F03 | Queue/repeat | Worker copied queue diverges from model; duplicates/repeat-one mishandled | IMPROVE | WEAK | HIGH | HIGH | main.rs, queue effects/finish_track; core, TrackBoundary; playback, copied tracks | Wrong next track/current state | Stable queue-entry IDs and live schedule contract | Yes | P0.3 |
| F04 | Crossfade | UI 5/10/15-s overlap exceeds real sink block limit; error fallback loses consumed head | IMPROVE | WEAK | HIGH | HIGH | playback, overlap send/fallback ~588–745; sink, 524288-byte bound; UI, SetCrossfade | Transition errors or skipped beginning | Chunk output, preserve samples on failure, test offered durations | Yes if exposed | P0.4 |
| F05 | Scanner/DB | Enqueue/barrier is treated as successful scan commit | IMPROVE | WEAK | HIGH | HIGH | library, DbCommand::Batch/Finish/Test, scan completion | Failed writes can precede successful pruning and false completion | Acknowledge writes; prune only complete committed scan | Yes | P0.5 |
| F06 | Persistence | Full queue checkpoint every 15 s; write/read size mismatch; silent restore default | IMPROVE | WEAK | HIGH | HIGH | main, periodic checkpoint/restore; core, 8-MiB restore vs unchecked checkpoint | Flash churn, UI stalls, lost session/settings | Dirty durable state, consistent bounds, migration/recovery reporting | Yes | P1.B |
| F07 | DB recovery | No corrupt-DB recovery; “Rebuild” is ordinary incremental scan | MISSING | MISSING | HIGH | HIGH | library, open/spawn; main, RebuildLibrary | Startup failure/crash budget; unusable library | Quarantine/recover explicitly without deleting music | Yes | P1.B |
| F08 | Collections | Album/artist play reduces to global index then queues whole library | IMPROVE | WEAK | HIGH | HIGH | UI, play_first_filtered, collection add; main, play_index | Collection actions do something different from their label | Preserve collection/entry intent through Effect | Yes | P0.3 |
| F09 | Format discovery | AIFF/APE/WavPack decoders exist but scanner ignores files | PARTIAL | PARTIAL | MEDIUM | HIGH | library, supported; defconfig | Music invisibly absent | Align discovery, tests and advertised format list | Conditional format claim | P1.A |
| F10 | High resolution | S32/high-rate wired path absent; current output S16/44100 | PARTIAL | PARTIAL | HIGH | HIGH | AFE, params/DAIs; profile; trace | 24/96 source precision/rate reduced | Truthful claim now; separate authorized hardware milestone later | Yes for high-res claim, no for scoped S16 alpha | P1 claims / P2 implementation |
| F11 | Gapless/seek | Limited FLAC continuity proof; seek label differs from sample position | PARTIAL | WEAK | HIGH | HIGH | media, rb_media_seek/decode_more; playback tests; probe | Imprecise resume/seek; uncertain lossy transitions | Waveform/count/delay tests and exact transition clock contract | Yes for promised behavior | P1.A |
| F12 | RG/EQ | Limiter auto-level changes headroom; weak tag validation; EQ toggle has no band configuration | PARTIAL | PARTIAL | MEDIUM | HIGH | media, make_filter_graph/gain_tag; pinned af_alimiter.c; settings, empty eq_bands | Unexpected level; misleading EQ On | Explicit gain/headroom policy and truthful minimal EQ interface | Gain behavior yes; EQ conditional | P1.A / P2 |
| F13 | Build | Clean cached-input acquisition/paired source/reused-payload identity not closed | IMPROVE | WEAK | HIGH | HIGH | build.py, ffmpeg9.py, reborn.mk; retained reuse receipt | Others cannot reliably reproduce intended release | One tested paired recipe and embedded payload provenance | Yes | P1.D |
| F14 | Manifest/docs | Validator requires obsolete application-unimplemented statement | IMPROVE | WEAK | HIGH | HIGH | package.py, manifest.application; validate.py, validate_manifest; Linux README | False release identity/capability contract | Update producer/validator/docs together after audit | Yes | P1.D |
| F15 | Distribution | Firmware/assets/fixtures/notices/source bundle not public-release closed | MISSING | MISSING | HIGH | HIGH | package source notice, reborn install, fixture rights, font licenses | Unresolved redistribution obligations | Actual image inventory, rights decision and compliance bundle | Yes for public binaries | P1.D |
| F16 | Power/charging | Current pack/source/recovery envelope lacks current physical closure; normal low-battery policy absent | UNKNOWN | UNKNOWN | HIGH | HIGH | charger, policy, 70-mA result, app power; commit 3dc0a8b | Safety/data-loss uncertainty | Freeze scope; current independent measurements and low-battery plan | Yes | P1.C; immediate scope gate |
| F17 | Suspend | GPU01 recorded failed entry/CPU3 timeout; correction not physical resume proof | PARTIAL | PARTIAL | HIGH | HIGH | GPU01, SPM, suspend helper | Sleep/wake reliability and battery claims uncertain | Same-boot current-image scoped regression | Conditional deep-sleep feature | P1.C |
| F18 | Wi-Fi | Scanning proven; association/IP/DNS/reconnect/coexistence not qualified | PARTIAL | PARTIAL | HIGH | HIGH | adapter, service, physical scans | May not provide usable network despite scan/connected UI | Define WPA2 subset; end-to-end connection/failure tests | Conditional Wi-Fi release feature | P1 connectivity |
| F19 | Bluetooth | SBC-only build; peer audio unproven; trust/reconnect seam and AVRCP missing | PARTIAL | PARTIAL | HIGH | HIGH | BT adapter, reconnect, generated BlueALSA config, physical inspection | Headset onboarding/reconnect/control uncertain | Fresh pairing/SBC proof; bridge trust/control policy; no optional-codec claims | Conditional Bluetooth release feature | P1 connectivity |
| F20 | Input | Timing before queued release, no SYN_DROPPED/reopen recovery, screen-off long-repeat inconsistency | IMPROVE | WEAK | MEDIUM | HIGH | input.rs, poll_at, ActionRouter::route | Wrong long action after stalls; stale keys | Backlog/timestamp/loss-aware tests and narrow correction | Yes for reliable ordinary controls | P1.B |
| F21 | UI truth/metadata | Literal 128-GB internal storage, SD-ID mismatch, route loss, incomplete glyph mapping | IMPROVE | WEAK | MEDIUM | HIGH | screens, storage text; UI, SD/back/catalog; components, glyph_index | Misleading status; unusable metadata/context | Bind real data/identities; preserve route; bounded correct glyph support | Truth/control yes; broad glyph support alpha limitation possible | P1.B / P2 |
| F22 | Graphics recovery | Real renderer exists; restart/pageflip-fault/current suspend recovery unqualified | UNKNOWN | UNKNOWN | MEDIUM | MEDIUM | native graphics, wait_flip/close/present; main, one recreation attempt | Black/stalled screen after fault | Physical lifetime/restart/screen-off tests; keep KMS architecture | Yes for offered recovery/sleep | P1.C / P2 |
| F23 | Scale | 20k first-page cap, full Track cloning and row rebuilding | IMPROVE | WEAK | MEDIUM | HIGH | main, library clone/query limit; library, Existing limit/count; UI rows | Large-library lag/RAM, hidden later records | Measure 10k/20k then bounded query-backed views/IDs | Conditional scale promise | P2 |
| F24 | Artwork | Preopened next art is published early; cache hits rewritten; sidecar invalidation absent | IMPROVE | WEAK | MEDIUM | HIGH | playback, open_next/publish_artwork | Wrong/stale art and avoidable writes | Entry-aware publish, read-only cache hit, source-aware invalidation | Current-track correctness yes | P1.B / P2 |
| F25 | Observability | Cumulative packet counter re-added; stale open-time metadata; statuses overstate completion | IMPROVE | WEAK | LOW | HIGH | playback, packet metric/audio_state; media metadata; F05 | Misdiagnosis from attractive but inaccurate telemetry | Correct semantics and test deltas before adding metrics | Not independently | P2 |
| F26 | Storage | ext4/vfat only; mount errors swallowed; hot removal/full disk/power loss not qualified | PARTIAL | PARTIAL | HIGH | HIGH | mount helper, adapter, main | Missing cards, interrupted playback, recovery uncertainty | Explicit filesystem policy, error propagation and failure qualification | Yes for advertised removable storage | P1.B |
| F27 | Validation | Historical receipts and permissive baselines leave current integration holes | IMPROVE | WEAK | HIGH | HIGH | ARM receipt, tooling tests, fixture tests, test review | False confidence in releases | Exact-source receipts plus tests at real failure boundaries | Yes | P0 tests / P1.D |
| F28 | Performance/power use | Wakeup/CPU/RAM budgets unmeasured for current app; polling/cloning costs evident | UNKNOWN | UNKNOWN | MEDIUM | HIGH | main, 15-ms loop; BT, 20-ms process; resource review | Battery/latency uncertainty | Measure representative target workloads, then simplify | Stable/battery claim | P2 |
| F29 | Board contracts | Vendor-derived timing/rail/memory assumptions remain qualification boundaries | UNKNOWN | UNKNOWN | MEDIUM | MEDIUM | DT, SPM, thermal, patch manifest | New hardware scope can invalidate known baseline | Preserve reservations/ownership; scoped regression before change | Conditional scope expansion | P2 / standing gate |
| F30 | Legacy | Old font asset/dependency metadata and pre-Reborn reference paths persist | DEAD/LEGACY | DEPRECATED / DEAD | LOW | HIGH | baseline docs, dependencies, current UI atlas references; debt | Maintainer confusion | Label authorities; delete only after usage/provenance review | No | P3 |
| F31 | Public install/recovery | Safe owner tooling still depends on choosing first-install vs preserving profile | IMPROVE | PARTIAL | HIGH | HIGH | layout/package, validator, init | Accidental USRDATA replacement by wrong valid profile | Independent public install/update/recovery exercise | Yes | P1.D |
| F32 | Security/privacy | Root native-media trust boundary; raw/platform diagnostics differ from sanitized app bundles | IMPROVE | IMPLEMENTED | MEDIUM | HIGH | control, sanitizer, collector, SSH | Malformed media is privileged; public log sharing needs care | Keep private defaults, review bundles and dependency maintenance; later scoped isolation | Sanitized release process yes; sandbox not automatic alpha requirement | P1.D / P2 |
Positive findings — preserve these
| ID | Subsystem | Finding | Classification | Evidence state | Severity | Confidence | Concrete evidence | User impact | Recommendation | Release blocker? | Suggested order |
|---|---|---|---|---|---|---|---|---|---|---|---|
| G01 | Boot/rescue | Strict identity/preflight/fsck and RAM rescue | GOOD | VERIFIED | LOW | HIGH | init, storage, production storage tests | Reduces destructive fallback risk | Preserve fail-closed selection | No; preserve | Stable |
| G02 | Storage packaging | Explicit protected targets and initialize-vs-preserve distinction | GOOD | VERIFIED | LOW | HIGH | layout, package/validation, readback tests | Preserves factory material/user data when correct workflow used | Keep allowlists/addressing contract | No; preserve | Stable |
| G03 | Platform layering | Real Linux subsystem/resource ownership, not app MMIO | GOOD | IMPLEMENTED | LOW | HIGH | AFE, MFG gate, CONSYS, DT | Maintainable driver/application boundary | Keep APIs/owners; qualify transitions | No; preserve | Stable |
| G04 | Graphics | Genuine Lima/GBM/EGL/GLES2/KMS hardware path | GOOD | VERIFIED | LOW | HIGH | renderer, GPU01 | Hardware accelerated UI with bounded scanout lifetime design | Keep; target recovery tests | No; preserve | Stable |
| G05 | Audio architecture | One FFmpeg processing authority shared by wired/BT | GOOD | IMPLEMENTED | LOW | HIGH | media, playback, generated component check | Avoids divergent decode/DSP implementations | Keep; correct lifetimes/transitions | No; preserve | Stable |
| G06 | DB ownership | Single SQLite worker, WAL, batches and source/path stable IDs | GOOD | VERIFIED | LOW | HIGH | library, cached library tests | Understandable concurrency and incremental metadata | Keep; add acknowledged scan outcome | No; preserve | Stable |
| G07 | Input boundary | Device-aware mapping and central semantic actions | GOOD | VERIFIED | LOW | HIGH | input, actual-name mapping/long/wheel tests | Avoids wheel/select collisions and duplicated mappings | Keep; fix backlog/loss semantics locally | No; preserve | Stable |
| G08 | UI/services | Draw is state-to-quads; services/hardware stay outside UI | GOOD | IMPLEMENTED | LOW | HIGH | UI, core effects, runtime | Later visual redesign need not rewrite hardware/playback | Preserve boundary, carry richer intent where needed | No; preserve | Stable |
| G09 | Atomic state primitive | Temp write + file sync + rename + directory sync | GOOD | VERIFIED | LOW | HIGH | core, atomic_write, restore tests | Better crash consistency than in-place overwrite | Keep primitive; change cadence/schema/error reporting | No; preserve | Stable |
| G10 | Diagnostics/private state | Bounded local control/logs plus private firmware/calibration handling | GOOD | VERIFIED | LOW | HIGH | observer, control, factory, storm/redaction/protocol tests | Useful physical debugging without unbounded flash logs | Keep bounds/privacy; correct status semantics | No; preserve | Stable |
VERIFIED on a GOOD row denotes the cited narrow source/test/historical evidence, not a whole-product safety or production certification. See validation tiers.