Resolve the k-d tree edge-scale 'proven wall': it was a 4-byte reader offset bug, found via eeu.mod's schema. Full tree traversal now reaches 100% of nodes/leaves; edge uses the disc-wide /100000-degree convention. Wire find_tile_for_coord_kdtree()/build_kdtree_bbox_index() into the viewer as the primary (structural, no-decompression) bbox source, with the cached content-decode path as fallback.
Document the geo-index disk cache: why the k-d tree wasn't swapped in yet, and the safe fix that was
Document confirmed-highway rendering using eeu.rd's cracked highway bit
Document eeu.iof nearby-streets lookup now wired into search-result jumps
Document road-search jump zoom now using eeu.rd's cracked delta_long/delta_lat
Migrate the map viewer's known-bugs and limitations lists from README
Completes the Map-Viewer-Tool.md migration: brings over the 3 bug/
limitation lists that sat after README's v1-v42 changelog (found
while building/testing v2, carried-over v1 bugs, and known v2
limitations) -- real, still-relevant reference material distinct from
the versioned changelog itself. Covers the _initial_scale() outlier
fix, the eeu.cty huge-bbox mitigation, the center-city-inclusion
guard, the Iasi bruteforce-anchor guard, the Sofia oscillation-spike
fix and its Turin multi-cluster follow-up redesign, and the explicit
list of known, accepted limitations (tile-boundary approximation,
road-importance-as-proxy, the still-uncracked topology split, mp0's
background-load window).
Migrate the map viewer's v28-v42 changelog from README to the wiki
Completes the migration flagged last turn: this page's own changelog
stopped at v27->v28 while README.md's ## 10 section had moved on to
v42. Brought over every entry in between, condensed to match this
page's own denser style but keeping every real number and validation
detail:
- v28->v29: a 2nd false-edge bug (the literal value 0 treated as a
real link id) surfaced once v27->v28's loosened mg3 cap stopped
masking it.
- v29->v32: the 3-step arc removing all disk extraction, reading tile
layers, then most remaining files, then POI.DB3 (via sqlite3's
deserialize()) directly out of the ISO.
- v32->v37: tile_ids_in_bbox() gains a content-bbox test, then a 3rd
native-hint test, then reorders mp0_ready around it -- then v36->v37
REVERTS the hint test as a real, measured 4.5-10x tile-count
regression, walked back honestly rather than patched.
- v35->v36 and v37->v38: a 3rd and 4th false-edge bug (a contaminated
3-point clique; a plain 2-point collision caught by a new gap-ratio+
distance conjunctive filter).
- v38->v40: the Sirius TravelLink POI layer and its real per-category
icons.
- v40->v41: live firmware screen-resolution reading, including a real
overclaim caught and corrected (a "156-entry" JAR central directory
that was actually fragmentary), and the splash-screen investigation
that ended in a strong, definitive negative result (rules out any
fixed-stride raw bitmap at any width, not just the formats tried).
- v41->v42: the 2 separate Address/POI search boxes.
Also fixed 2 stale pointers this migration exposed: the "Latest" tag
still said v22->v23, and the intro line still said "v1 -> v23+".
Migrate the theft-protection/C166 and Sirius/NAR research threads to the wiki
A consistency audit found 2 major real research threads that existed
in research/swl_5238_reader.py (and, for one part, README.md) but had
NEVER been migrated to the wiki at all, across the project's whole
history -- not just this session. Moving them in fully rather than
just linking out, per direction to consolidate documentation in the
wiki.
New page: Theft-Protection-and-the-CAN-Gateway.md
- GATEWAY.FLI's complete TheftProtection_* state machine (string
level): status dump fields, diagnostic override interface, ITD_*
function family, a separate CodingShadow_*/CONFIG_FKTID_* coding-
sync protocol.
- A real ST10/C166 disassembler (hn/c166-dis) found and validated --
no capstone support for this chip exists.
- A real string-to-code cross-reference: found the actual diagnostic
override dispatcher, a disassembly-confirmed 50-retry limit, and the
real 0x5A55/0xA5AA EEPROM signature scheme -- all at the instruction
level, not inferred from strings.
- 7 real techniques tried to find the dispatcher's own caller, all
negative, including one caught-and-refuted false positive (a linear
sweep desyncing into a data table) -- a real, general lesson about
this architecture recorded so it isn't hit again.
- TpPvVerifyPin's real location in FHDD6.FLI (~82.6MB, the VxWorks/BSP
region) and the real ClusterID vehicle-pairing security design --
correcting an error in this session's own earlier wiki edit, which
wrongly said TpPvVerifyPin was "not in FHDD6.FLI" when the research
file clearly places it there by file offset.
- Consolidated (moved, not duplicated) the serial-console confirmation
and wiring diagram already added to HOST-BSP-and-Vehicle-Coding.md
last turn -- that page now points here instead of repeating it.
New page: Sirius-TravelLink-and-North-America-Discs.md
- The first 2 real North-America SWL discs: real part-number
prefixes, a SetCoding byte-7 refinement (a real enum, not a bool),
and the first confirmed case of the testmode register explicitly
OFF.
- SIRIUS.DB3, a real Sirius TravelLink SQLite database -- the map
disc's own POI.DB3 coordinate decoder applies completely unchanged.
- CHANNELARTSCACHE/*.DDB, a complete real image format crack (2-plane
BGR555+mask, visually confirmed against 3 real Sirius channel
logos), with the filename-to-database key cross-checked 3 for 3.
- A clean IBOC/HD-Radio negative result across all 5 discs on hand,
and a 4th independent cross-build confirmation the 9MB+ FHDD6.FLI
disassembly wall is structural, not EU-specific.
Also: added the "other ECU images" first-pass findings (DAB.FLI's
real TI DSP identity) to the boot-chain page, and flagged (without
attempting a full migration) that Map-Viewer-Tool.md's own changelog
is stale relative to README's -- a separate, large body of app
changelog content, different in kind from these RE findings.
Document the max_edge_m per-layer fix for mg4/mg3
Document removal of the Hide decode garbage checkbox
Document the new Show tile boundaries checkbox
Document the temp-workdir leak fix (146GB reclaimed)
Map viewer: document v22->v23 (POI search)
Mirrors the main repo's README S10 "v22 -> v23" writeup: the unified
search box now also searches real POI names, completing the POI
feature arc (crack -> display -> icons -> click-to-identify -> search).
Map viewer: document v21->v22 (POI click-to-identify)
Mirrors the main repo's README S10 "v21 -> v22" writeup: right-clicking
a real POI icon now shows its name/category, checked first (wider pick
radius) and restricted to exactly the declutter-filtered set the last
redraw actually drew.
Map viewer: document v20->v21 (real POI category icons)
Mirrors the main repo's README S10 "v20 -> v21" writeup and the
research/poi_db_reader.py load_poi_icons() crack: real, standard PNG
icons throughout PoiPartition -> Image -> ImageBlob, a real grid-based
declutter fix found and verified on the actual ISO (a dense Sofia view
first rendered as an unreadable wall of icons, fixed and re-verified by
inspecting a real screenshot).
Map viewer: document v19->v20 (real POI display)
Mirrors the main repo's README S10 "v19 -> v20" writeup: real POI
display from EDB/POI/POI.DB3, unlocked by the same-session Coordinate
crack, zoom-gated by the disc's own real MapZoomLevel design, rendered
into the same rasterized bitmap as road dots.
Map viewer: document v18->v19 (perf fixes + nearby-streets feature)
Mirrors the main repo's README S10 "v18 -> v19" writeup: 2 measured
performance fixes (cached per-tile road-name matching, cached canvas
size in _to_canvas()) and the new eeu.iof/.il "nearby streets"
click-to-identify feature.
Add full file-format wiki: home status table + per-file detail pages
Documents every file on the CD_8555 map disc (db/ folder): byte layouts,
decode status (cracked/partially cracked/not examined), and validation
evidence for each, distilled from the main repo's README research log.
Covers the road graph (eeu.rd/.il/.iof/.typ), the road/city name-search
catalogs (rt/rl/prl, cty/cl/ct, ctr), the MAP_COMPRESSED tile format
(mp0/mg1-mg4) including tile geometry and the topology/adjacency table,
the eeuz.fea gazetteer, the FLAT_COMPRESSED container, ISO handling, and
a first-look triage of the 16 files never examined at all.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>