# Progress ## Milestones ### Hardware: SB-LoRa v0.1 - [x] Board designed, DRC clean, fab package generated - [x] Check part rotations in JLC preview (U1–U7, Y1, D1); corrections saved to the CPL - [x] Order PCBs + assembly from JLCPCB (5 boards, 2-layer, 1.6 mm, leaded HASL); tagged `hw-v0.1` - [ ] Boards received - [x] Order J1: Hirose IC1HA-68RD-1.27SH(51) is obsolete; 1 ordered as new-old stock from eBay ($24) - [ ] J1 order cancelled 2026-11-17; rev 1 gets a harvested or wired-on connector instead (README.md, "Without the Hirose part") - [ ] Order 2x 1x7 female headers - [ ] Buy 2–3 cheap used Springboard modules (spare connectors to harvest + slot-fit reference) - [ ] Bare board: continuity check, dry-fit in Visor slot - [ ] Fit J1, bench-power VSB at 3.3 V, check current and 14.7456 MHz on Y1 - [ ] In Visor: read/write SC16IS740 scratch register (SPR, 0x07) from Palm side ### Hardware: SB-LoRa rev 2 - [x] Plan (hardware/sb-lora/REV2.md): S29AL016J flash on CS0 so the app loads on insertion, SC16C650B UART on CS1 in place of the SC16IS740, XIAO and Wio side by side, LiPo pads - [x] Netlist, footprints and LCSC numbers; 4-layer layout routed, DRC 0 unconnected, verify.py passes - [x] Choose J1: Robinson Nugent / 3M MCS-068SMA1-TR30, female straddle receptacle; ordered from eBay 2026-11-19 - [x] J1 laid out as a straddle (pins 1-34 on F.Cu, 35-68 on B.Cu opposite, 1.27 mm); rerouted, DRC 0 unconnected, verify.py passes; row position, pad size and thickness estimated - [ ] Measure J1 (row gap, tail positions, pad size, body), set `J1_ROW`, `J1_PAD` and `BOARD_T` in build_board.py, rebuild - [ ] Check the XIAO and Wio footprints against the boards at 1:1; check rotations in JLC's preview - [ ] Order; tag `hw-v0.2` - [x] Draft case (hardware/sb-lora/rev2/case/): two-part OpenSCAD shell for FDM, in-slot part from nkorth's CC0 slot cover, bump for the radio boards, LiPo and an FPC antenna; not printed - [ ] Print nkorth's blank and check it fits the slot - [ ] Measure the connector's contact height in the slot, J1's body, and the radio and LiPo heights; set them in sb_lora_case.scad and print the case ### Firmware - [x] Apply `firmware/meshcore-xiao-nrf52-uart-companion.patch` to MeshCore, build `Xiao_nrf52_companion_radio_uart` (`firmware/Makefile`, MeshCore a366955c, GCC 14 on Apple Silicon) - [x] Flashed to the XIAO; answers `CMD_DEVICE_QUERY` over USB with DEVICE_INFO (v1.17.1) - [x] Bench-test the D6/D7 UART: FT232R at 3.3 V, 115200 8N1, DEVICE_INFO identical to the USB reply - [ ] Plug stack into carrier; `CMD_DEVICE_QUERY` gets a DEVICE_INFO reply ### Palm client - [x] Toolchain set up: prc-tools-remix 2.3.6 + PilRC + Palm OS 3.5 SDK in Docker (`palm-client/Dockerfile`) - [x] Scaffold: builds `visormesh.prc` (main form + "Query radio" → CMD_DEVICE_QUERY); protocol framing with host tests - [x] USB link from the dev Mac: pilot-link (desrod @6d696653 + `tools/pilot-link-darwin-usb-open-retry.patch`) lists the Visor's databases - [x] `make install` puts VisorMesh on the Visor; app launches and shows the main form on the Visor Neo - [ ] Emulator: CloudpilotEmu with a Palm OS 3.5 ROM (the Visor isn't emulated); optionally POSE + a Visor Neo ROM in a Windows VM - [x] Original Handspring SDK headers (Springboard `HsCard*` APIs; the PalmOne `HsExt.h` lacks them). The calls used are declared in `src/hs_card.h`; the SDK is in `docs/reference/Handspring_SDK.zip` - [x] Bus SPI routines and SC16IS740 driver (`sb_spi.c`, `transport_sb.c`); host-tested against a simulated carrier, not on hardware - [x] Companion protocol: framing, builders (DEVICE_QUERY, APP_START, GET_CONTACTS, SYNC_NEXT_MESSAGE), parsers (SELF_INFO, contacts, v3 contact/channel messages); host tests from real captures + MeshCore layouts - [x] USB bench link: app (`HsUd`) → Visor USB cradle → `tools/visor_bridge.py` → FTDI → XIAO D6/D7. Query radio shows `Reply 0x000D, 82 bytes` on the Visor - [x] Connect screen: DEVICE_QUERY + APP_START via session.c; shows node, freq, BW/SF/CR, TX power, firmware (verified on the Neo over the USB bridge) - [x] Node Settings screen: name, preset/radio params, TX power, path hash. Verified on the Neo: SET_RADIO_PARAMS + SET_PATH_HASH_MODE both got RESP_OK and a read-back over the XIAO's USB matches - [x] Contact list: GET_CONTACTS into a Palm database + contacts screen - [x] Messages: store, sync on PUSH_MSG_WAITING, send with ack tracking (built; to verify on the Visor) - [x] MeshTerm-style UI: small 4x6 font, Home (conversations), Conversation, Contacts, Node (built; to verify) - [x] Split the code into segments: protocol.c + ui/text.c in vmcode2 (seg 1 24.3 KB, seg 2 4.9 KB); rules in src/sections.h (to verify on the Visor) - [ ] UI: contacts, DMs, channels - [x] Install method confirmed: HotSync over the USB cradle from the dev Mac (`make install`) - [x] Notification sounds per event kind, mute channels, larger text option (5x8 font) - [x] Flood scopes (regions): manage the list of scopes, set the default scope, and set a scope per channel; discover from neighbouring repeaters. MeshCore v8+ has CMD_SET_FLOOD_SCOPE_KEY (54), CMD_SET_DEFAULT_FLOOD_SCOPE (63) / CMD_GET_DEFAULT_FLOOD_SCOPE (64) - [ ] Map: clip polygons that leave the box (the lower Delaware shows only its banks); optional LZ tiles to halve the size; topo contours - [ ] Region discovery beyond neighbours: keep each contact's out_path and send the reversed path as the reply path ## Open questions - Which CF adapter is it, and does its ROM include an app installer? - Is the J1 footprint correct against the Hirose drawing? (Only checked against springSD.) - Which Handspring offset variant (−0.3 / 0 / +0.3 / +0.6 mm) is the (51)? This matters if a harvested connector replaces it. - Long-term J1 supply: no current-production card-side part found (Amphenol, Hirose, TE and Assmann parts are obsolete or host-side). Rev 2 uses eBay surplus MCS-068SMA1, so buy enough for every board. Does a PCMCIA card-side straddle receptacle seat fully in the Springboard slot? (The keying is in the case; untested.) - What board thickness does the MCS-068SMA1's row gap take? - Does the extension clear the Visor body? Is the slot-opening alignment right? ## Log ### 2026-10-30 - Planning done in claude.ai. Chose a custom Springboard module (SB-LoRa) over cradle serial/IR. - SB-LoRa v0.1 fab package generated. Repo set up and files organized. - Crystal Y1 swapped to SX3B14.7456F2010F30 (C2901666, 20 pF load) because C6264361 was out of stock. - JLC placement preview showed U1, U2, U4 a quarter-turn off and U3, U5–U7 a half-turn off. Fixed in the preview and in the CPL (270 / 180). - Chose leaded HASL over ENIG (+$18). JLC does the fine-pitch assembly, so pad flatness isn't critical. - PCB + assembly order placed with JLCPCB. `hw-v0.1` tag moved to the ordered commit. ### 2026-11-01 - J1 (IC1HA-68RD-1.27SH(51)) found to be obsolete at DigiKey and Mouser. Ordered 1 NOS from eBay at $24. - Removed the unverified XKB/LCSC substitute from the hardware README (it's a vertical SMD socket, not a straddle). - Starting Palm client development. - macOS USB sync to the Visor works with pilot-link from desrod's repo. Fixed a race where macOS holds the Visor briefly when it appears, so pilot-link's single open attempt failed with exclusive access. The retry patch is saved in `palm-client/tools/`. - First on-device run: VisorMesh v0.1 installed with `make install` and launches on the Visor Neo. - Firmware builds and runs. Fixed the patch's I2C override (`-U` flags broke Wire under PlatformIO; now `build_unflags`). Pinned native GCC 14 on Apple Silicon. Enabled the USB companion link alongside the UART. The XIAO answers DEVICE_QUERY over USB, and that real frame is now a Palm host test. - UART path verified: FT232R on D6/D7 at 115200 8N1 gets the same DEVICE_INFO as USB. The radio side of SB-LoRa is proven; what remains is the carrier board (SC16IS740 + bus) and the Palm transport for it. - Protocol layer for the read-only commands, built against MeshCore's MyMesh.cpp and captured replies (DEVICE_QUERY, APP_START → SELF_INFO, GET_CONTACTS with 0 contacts, SYNC → NO_MORE_MESSAGES). LE fields are assembled byte by byte for the big-endian 68k. The radio is on MeshCore factory defaults (869.618 MHz / 62.5 kHz / SF8, 22 dBm, no name). - **First end-to-end round trip from the Visor:** DEVICE_QUERY from VisorMesh on the Neo went over USB, through the Mac bridge and the FTDI adapter, into the XIAO's UART, and DEVICE_INFO came back and decoded on the Visor. `HsUd` turned out to use the Visor's HOTSYNC USB port (2); the endpoints stall around open/close and need clear-halt. - Connect screen verified on the Neo: shows the radio's real SELF_INFO/DEVICE_INFO (9B1E4F92, 869.618 MHz, BW 62.5 SF8 CR5, 22 dBm, v1.17.1). Fixed a hand-typed DEVICE_INFO fixture that had two bytes shifted. - Node Settings verified from the Visor: radio now 919.500 MHz / BW 500 / SF10 / CR5, 2-byte path hash (confirmed independently over the XIAO's USB). Seen in the bridge log: the Visor's USB link dropped while the Settings form was open and re-attached on the next send; it worked, but commands only try once, so this needs a closer look. - Reception was dead until the XIAO rebooted: SET_RADIO_PARAMS saves and reports the new settings, but the receiver only retunes on a restart. Settings now offers to reboot the radio (CMD_REBOOT) after a radio change and reconnects. The radio then heard the user's contact. - MeshTerm-style UI built: generated 4x6 font (PilRC's FONT compiler writes a corrupt NFNT, so gen_font.py writes it), rowlist, text layout (host-tested), message store and sync, Home/Conversation/Contacts/Node screens. Font verified readable on the Neo. - Transcript wrapped after ~4 characters on the Visor: prc-tools puts `const` data in the code section, but another file reaching an `extern const` table addresses it A5-relative as data, so it read garbage widths. Shared tables (font widths, presets) are now non-const, set in their generators. Host tests can't catch this. Rule: never share an `extern const` table across files in palm-client. - Times were wrong: Palm OS 3.5's clock is local time with no zone, and it was used as UTC (and the Visor hadn't moved off DST). That made contact ages read 'now', and clock sync pushed a wrong time into the radio, which can only move forward. Now everything stored and sent is UTC via a UTC-offset app preference (Settings), the display converts back, and names are converted from UTF-8. - Clock saga, resolved: the Neo's prefMinutesWestOfGMT was 1140 (19 h), so the default UTC offset was nonsense. The Visor's date was Oct 31 2026, then mistyped as Oct 1 2027, and each was copied into the radio, whose clock only moves forward. MeshCore restores its clock at boot from the newest contact "lastmod", so a contact heard during the 2027 window (TGNAB-01) kept bringing 2027 back. Fixed by rewriting that contact's lastmod (CMD_ADD_UPDATE_CONTACT with the optional last_mod field) and rebooting. The app now ignores system zones past +-14 h, only corrects a set radio clock by under 7 days (otherwise Node shows MISMATCH), and shows dates on the Node clock line. Radio within 7 s of real UTC. - Mentions draw as chips and tap-to-reply works on the Neo. Two code segments verified on the device. - Channel management verified on the Neo: Add Channel "#test" put key 9cd8fcf22a47333b591d96a2b848b73f in slot 1 (matches companion_protocol.md). Reconnect after sleep verified: power off ~15 s, the app reconnected and synced on wake by itself. - Docker Desktop's amd64 emulation hung once (containers never started); make install now only uses Docker when sources changed. - Verified on the Neo: two-line compose field (scrolls, redraws after the on-screen keyboard), Unicode symbols/emoji smiley, line breaks in messages, beep for DMs/mentions, "new" divider. Segments: 1 = 7.8 KB, 2 = 27.1 KB (the next big feature should go in a third segment); tools/check_sections.py guards the rules on every build. - 2026-11-02: Verified on the Neo against a PhillyMesh repeater: login, admin console (Read settings, typed CLI commands, Status 0x87 and Telemetry 0x8B decoded), contact detail with signal history, Add Contact, Reset Path. Segments: 9.1 / 27.7 / 12.4 KB. Upstream issues written up in docs/upstream-issues.md (none filed yet). - 2026-11-02: SB-LoRa transport written ahead of the boards: bus-SPI bit-bang over CS0/CS1 (`sb_spi.c`), SC16IS740 setup for 115200 8N1, burst FIFO access, sleep, and a "Slot" link on the Node screen (`transport_sb.c`). `tests/test_sblora.c` checks it against a simulated carrier (latches, inverter, SC16IS740). Unknowns for bring-up: whether connector A0 is CPU A1 (the probe tries both mappings via the scratch register), the hard-coded csSlot0 base 0x28000000 (no Handspring SDK for HsCardAttrGet), and RX overruns while the app isn't polling (no IRQ handler yet). Segments: 9.6 / 27.7 / 13.1 KB. - 2026-11-02: Slot with an empty slot reset the Visor (bus error at csSlot0), and since Slot was saved and Home auto-connects, every launch reset it. Slot is no longer restored at launch, and Connect now checks `HsCardAttrGet(hsCardAttrHwInstalled)` and takes the chip-select base and size from the OS before any bus access. Verified on the Neo: empty slot gives "No module in the slot." with no reset. Node-screen errors shortened to its 22-character labels. - 2026-11-02: Notification sounds (Options > Sounds): separate tunes for direct messages, mentions, room posts, private / hashtag / Public channels and new contacts (PUSH_NEW_ADVERT), each on or off, at the alarm volume. One sound per poll, by priority; none for the conversation on screen. Verified on the Neo. Segment 3 is 13.8 KB. - 2026-11-02: SB-LoRa receive is interrupt-driven: a slot IRQ* handler (hsCardAttrIntHandler) drains the SC16IS740 RX FIFO (IER RHR, trigger 8 or timeout) into a 1 KB ring passed as cardParam. The handler path (SbDrainRx and the bus core) is in segment 1 with no A5 references, checked in the disassembly, so it's safe in interrupt context and while the Visor sleeps. The app masks the slot interrupt around its own SPI transactions, and falls back to polling if the handler won't install. Host tests cover draining, ring wrap and overflow; untested on hardware. - 2026-11-02: Mute Channel (Options menu; by channel key; mentions still sound; open-circle unread badge, left out of the title total) and Text size Small/Large on the new Preferences screen (5x8 misc-fixed via tools/pcf2bdf.py from XQuartz's PCF; gen_font.py builds both fonts). Verified on the Neo. Segments: 11.1 / 28.0 / 14.7 KB. - 2026-11-02: Flood scopes: a list of scope names (Options > Flood Scopes), the radio's default (Set Default / No Default), a per-channel choice (Options > Channel Scope: radio default, none, or a named scope), and Discover, which asks neighbouring repeaters for their regions (SEND_ANON_REQ type 0x01). The session sets the scope override before each flood send and resets it on disconnect. Keys and frames are host-tested (tests/test_scopes.c; the key for "test" matches the radio's #test channel key). Verified on the Neo. Discover only reaches 0-hop repeaters for now. - 2026-11-02: Map screen (Options > Map): an OpenStreetMap basemap of the Philadelphia metro in 4 greys with the contacts that advertise a position plotted on it (no position, not shown). `tools/make_map.py` downloads OSM roads, rail, water and parks (Overpass, cached in tools/map-cache), draws zooms 10-12 for the metro and 13 with residential streets for ~16 km around Center City, and writes build/VisorMeshMap.pdb (290 PackBits tiles, 1,045 KB); `make map`, `make install-map`. Zoom with -/+ or the hard buttons, drag to pan, tap a node for its info, tap again to open it. Fixed on the device: a long node name overflowed the info bar (reset), and the attribution overlapped the buttons. Verified on the Neo. Segment 3 is 24.9 KB. - 2026-11-02: Emoji: 12 common ones have their own glyph in both fonts (thumbs up, tears of joy, heart, crying, folded hands, waving hand, fire, party popper, thinking, eyes, antenna, 100), with close variants mapped onto them and skin tones dropped; the rest stay the generic smiley. Tap an emoji in a message for a 32x32 picture and its name (resources/emoji.bin from gen_font.py). Verified on the Neo. Segments: 12.2 / 28.8 / 25.7 KB; vmcode2 is nearly full. - 2026-11-02: Fourth code segment (vmcode4: map and flood scopes). Segments: 12.2 / 28.8 / 15.6 / 10.2 KB. Home's title status erases back to where the previous text started (the large font left "Co" of "Connecting..." behind). Verified on the Neo. - 2026-11-03: Check for messages while the app is closed (Preferences > Check when closed: 5/15/30/60 min). An Alarm Manager alarm launches the app without globals; it sublaunches itself with sysAppLaunchFlagNewGlobals to connect without the channel/contact refresh, sync what the radio queued and play the sound, then keeps the alarm for display: an alert with the count per conversation ("3 new messages", "#test: 2") and Open. Found on the device: SysCurAppDatabase names the launched app itself in the alarm launch, so every check thought the app was open (now a feature flag); the channel list wasn't open in the check, so the alert showed "Channel N", channel messages sounded as private and mutes were ignored. Preferences shows the last check's result and the next time. Verified on the Neo over the USB bridge. Also: menus split into Go / Channel / Options per screen, and the app icon from meshcore-visor-hair-1.png (tools/make_icon.sh). - 2026-11-03: Tap and hold a channel on Home (half a second, pen within 3 px; the row unhighlights as the cue to lift) for a popup of Mute/Unmute, Channel Scope and Leave Channel. RowList has an optional hold callback for other lists later. Verified on the Neo. Segments: 13.6 / 29.3 / 15.8 / 11.3 KB. - 2026-11-03: Tap and hold extended: DMs get Pin/Unpin, Contact Info, Reset Path and Remove Contact (shared with the contact screen); channels get Pin/Unpin too. Pinned conversations move to a Pinned section at the top of Home (channels in slot order, then DMs most recent first); empty sections are hidden. Pins are kept by key in prefs (pins.c, segment 4). The popup now flips above rows near the bottom of the list. Verified on the Neo. - 2026-11-03: Screenshots: Options > Screenshot (or Screenshot in 5 s, to catch an open menu or popup) copies the screen into the VisorMeshShots database with a click; `make shots` fetches it over HotSync and tools/shots.py writes 16-grey PNGs to build/shots, named by capture time. Verified on the Neo. - 2026-11-04: Fixes from a code review. Channel messages now carry the channel's secret prefix, so their history follows the channel when its slot changes, is set aside when it's left and comes back when it's added again (before, the next channel in that slot inherited it). The "new" divider is tracked by record unique ID, so it no longer drifts once the 500-message store starts dropping old ones. Reply waits can't pass a negative timeout to the serial manager. Standard Edit menu (Undo, Cut, Copy, Paste, Select All, Keyboard, Graffiti Help) on every screen with a text field; app shortcuts moved off C, P and S (Contacts T, Serial Ports I, Screenshot Z). Verified on the Neo. Segments: 13.8 / 30.1 / 15.6 / 12.0 KB. - 2026-11-04: Reconnects (after sleep) download only the contacts changed since the last download (GET_CONTACTS with since); the first connect of a run is still a full refresh. The DM ack timeout no longer scans the 500-message store every poll, only once the earliest deadline can have passed. Closed-app checks that find no radio space out (10, 20, 40, then 60 min at a 5-min interval), and reaching the radio restores the interval. Verified on the Neo. Segments: 14.1 / 30.2 / 15.6 / 12.3 KB. - 2026-11-04: A connect gives way to the user: a tap or key press stops it within 0.1 s before the radio has answered (offline, "Connect stopped."), and after the handshake skips the refresh steps left, between requests so no reply is lost. The tap is then handled as usual. Closed-app checks give way the same way to the app in front. Verified on the Neo. - 2026-11-04: The rest of the review. Mentions don't wrap across lines; an SNR that doesn't fit after a wrapped message goes in its last line's time column; "tap for earlier messages" at the top of a conversation shows 40 more at a time. Contacts shows the route only when it's known, so names get the room. Options > About VisorMesh shows the app version and the radio's firmware. One DbOpen for the three stores (in vmcode4). Verified on the Neo. Segments: 14.7 / 30.3 / 15.6 / 12.5 KB. - 2026-11-05: Connection state in conversations' title too (the title is cut with an ellipsis to make room); tapping "offline", or Options > Connect, connects. A command the radio doesn't answer closes the session (it stayed "online" with every command failing), and after a quiet minute a time query checks the link. Reconnects after a wake or a lost link run in steps on idle events, so the app stays usable ("connecting" in the title), and are retried 2, 4 ... 256 s apart; before, the reconnect on wake often ran before the USB link was back and was never tried again. Found on the device with a timing readout: with nothing on the host reading (bridge stopped), SrmSend over USB blocks for 10 s whatever the CTS timeout is set to, and CTS is on only while the host reads, so USB sends now wait for CTS instead. Verified on the Neo. Segments: 16.3 / 30.2 / 16.1 / 12.5 KB. - 2026-11-06: Radio Stats (Go > Radio Stats, or Stats on Node): battery, storage, uptime and queue, noise floor and the last packet's signal, airtime and packet counts from GET_BATT_AND_STORAGE and GET_STATS, with a noise floor graph sampled every 5 s. Fixed on the device: the lines ran off the right edge, so they're shorter, too-long lines are cut, and the graph takes the height left. Verified on the Neo. The other features from MeshCore One (trace, auto retry, heard packets, hashtags, mentions, blocking, stale contacts, beaming, route editing, repeater neighbours and regions) are untested. A fifth code segment (vmcode5); Node, Settings, Font Preview and Add Channel moved to vmcode4. Segments: 18.8 / 23.6 / 22.2 / 25.1 / 19.5 KB. - 2026-11-06: A contact's More list shows what suits its type (Admin Console, Trace / Route, Ping, Reset Path, Block/Unblock, Beam, Remove); its choices verified on the Neo. Fixed on the device: More ran past the right edge, so the contact screen's buttons have fixed widths. - 2026-11-06: Heard Packets (Go > Heard Packets): the last 40 packets the radio heard (RX log pushes), newest first, with kind, route, SNR and the advert's node or last relay; All / Chat / Adverts filters and a tap for the details (time, SNR and RSSI, the hops by name). Kept in memory while the app runs. Verified on the Neo. - 2026-11-06: Typing @ in a conversation (or Options > Insert Mention) lists the recent senders there, then the companions in the contacts, and inserts @[Name]. Verified on the Neo. - 2026-11-06: Block a sender (a contact's More, or hold their message > Block Sender): kept in prefs by key and by name, listed in Contacts > Blocked Senders to unblock. Blocking and unblocking verified on the Neo; dropping a blocked sender's new messages as they sync is untested. - 2026-11-06: Contacts > Remove Stale Contacts: removes from the radio the contacts it hasn't heard an advert from in 2 weeks, a month or 3 months, after saying how many; it keeps those with messages on the Visor, those never heard (added by key) and those we're logged into. The choice and the count verified on the Neo; the removal itself is untested. Fixed on the device: the alert's four buttons didn't fit (now 2 wk / 1 mo / 3 mo / Cancel). - 2026-11-06: Ping a repeater (its contact's More > Ping): a trace whose path is just the repeater (SEND_TRACE_PATH, 2-byte hashes as the radio's path hash size), answered by PUSH_TRACE_DATA with the tag sent, showing the time and the SNR each way. Verified on the Neo against a PhillyMesh repeater (78e6: +9 dB there, +8.25 dB back, about 1 s); 219b gave no answer to a ping; the trace below shows it's only reachable through 78e6. - 2026-11-06: Trace Path (a contact's More > Trace / Route, or Go > Trace Path): hops filled from the contact's route (GET_CONTACT_BY_KEY) or typed / added from the repeaters, Round trip sends it out and back (A B A). Verified on the Neo: 78e6 219b 78e6 answered in about 1 s with every leg's SNR (+9, -16, +7.5 and +8.75 dB). Set Route (rewriting a contact's route on the radio) is untested. - 2026-11-06: Repeater neighbours on the Map: the admin console's "neighbors" reply (key prefix:secs ago:SNR per line) is kept for 30 minutes, and the Map draws a line from the repeater to each neighbour with a position, solid for an SNR of 0 dB or better, grey below. Verified on the Neo with 78e6. - 2026-11-06: The Map's own Map menu: Basemap On/Off (M) and Neighbour Lines On/Off (L), remembered with the view. Verified on the Neo. - 2026-11-06: Map: the street-level zoom covers about 40 x 31 km (the Main Line to Cherry Hill, Chester to Cheltenham; the map is 2.0 MB). Ardmore was blank there (outside the old ~16 km area); now it shows streets. Verified on the Neo. Outside that area the next zoom out is shown enlarged instead of blank tiles (untested). - 2026-11-08: Text that ran off the right edge now wraps: Radio Stats' "Connect to the radio first" and the SB-LoRa Test's messages (new DrawWrap), and Add Channel's note (three lines). Verified on the Neo. tools/check_ui.py now runs on every build: it checks form objects and labels against their form, alert buttons against Palm OS's layout, and literal DrawCut text against the large font. It found the Add Channel and Add Contact notes and a Font Preview line. Not yet checked on the device: Add Contact's note, the SB-LoRa Test with the slot empty, and the command retry (frames that didn't go are sent again; commands that are safe to repeat get a second try). Segments: 19.0 / 23.7 / 22.4 / 25.9 / 22.2 KB. - 2026-11-08: Node's clock lines and Radio Stats fit the screen. Verified on the Neo. They're built in screen_fmt.c now, and tests/test_screens.c checks their longest values in the large font. That test is what found them: the Node clock ran off the edge even in the small font, and Radio Stats lines with big counts were too long (now they wrap). Also new on every build: tools/check_source.py (function-level static const in extra segments, menu items no code handles, clashing shortcuts, forms with no handler, repeated IDs). Not yet checked on the device: the contact summary ("Room server" is now "Room") and Trace's result line. Segments: 19.0 / 23.7 / 22.4 / 25.6 / 24.1 KB. - 2026-11-08: The contact screen's login status ("Logging in over the mesh...") fits; its status line is now cut rather than running off the edge. Verified on the Neo. check_ui.py missed it because the text went through the screen's own Status() helper, so it now also measures literals passed to such helpers, and fails on a new one until its width is listed. - 2026-11-08: Verified on the Neo: Trace's result line fits ("Back in 1.2 s, N hops"); the command retry (with the bridge stopped a command goes offline after a few seconds, and after a reconnect everything works as before); the SB-LoRa Test with the slot empty says to insert the module and doesn't touch the bus (no reset); Add Contact's three-line note fits. The SB-LoRa Test's bus steps (mapping, registers, loopback, interrupt, radio) are untested until the board arrives. - 2026-11-08: Slot link restored at launch only while the OS reports a module. Verified on the Neo with the slot empty: Connect on Slot says "No module in the slot", and a relaunch doesn't reset the Visor and comes back on USB. Also new, untested until the board arrives: the radio's "message waiting" push wakes the Visor on the Slot link (open app: sync and sound, screen off; closed app with Check when closed on: a copy of the handler left in system-owned memory starts the closed-app check at once), checked on the host with the simulated module; and hardware/sb-lora/BRINGUP.md, the step-by-step bring-up plan. Segments: 20.1 / 23.7 / 22.4 / 25.6 / 24.1 KB. - 2026-11-08: Verified on the Neo: the three overlaps check_ui.py's new overlap rule found are fixed (Node "Version:", Preferences' Check when closed popup, Flood Scopes' default), and Add Channel's New Key gives a private channel a random key from the entropy pool (pen and key timing, the radio's packet readings; kept between launches in an unsaved preference). Adding a channel with it is untested (it writes to the radio). Also new, host-only: the request retry rules in request.c with tests/test_request.c and check_source.py's try-once rule, and make replay over the bridge's full-frame capture. Segments: 20.8 / 23.7 / 22.4 / 26.0 / 24.8 KB. - 2026-11-09: SB-LoRa rev 2 planned in hardware/sb-lora/REV2.md: a flash on CS0 so the app loads when the module goes in (the installer copies it to RAM, since the transport's interrupt handler can't live on a removable module), a 16550-style UART on CS1 in place of the SC16IS740 SPI bridge, and the XIAO and Wio side by side to make the module flat. Netlist, footprints and LCSC numbers done. - 2026-11-11: Release workflow: a v* tag builds and tests the app and publishes visormesh-.prc with SHA256SUMS. The toolchain install is shared with the Dockerfile (tools/install_toolchain.sh). - 2026-11-17: The Hirose IC1HA order for rev 1's J1 was cancelled. README.md lists the stopgaps for rev 1 (harvested straddle part, donor connector on wires). Rev 2 laid out: 4 layers, routed with Freerouting 1.9.0, DRC 0 unconnected, verify.py passes; J1 left as a placeholder. - 2026-11-19: Rev 2 J1 chosen: Robinson Nugent / 3M MCS-068SMA1-TR30, a female MEMPAK straddle receptacle at 1.27 mm with the rows directly opposite each other, ordered as eBay surplus. Ruled out as obsolete, host-side or the wrong part: Amphenol 10014744-007T, 71299-050CALF and 7111S1615A01LF (a SIM tray), Hirose IC1F, IC9 and IC1K, TE 5535657, Assmann A-PCMCIA-68 (male THT), and an AliExpress female part (photo and listing brands disagree; $39 shipping). The single-row MCS-068SSA1 would suit a single-sided board but no stock turned up. J1 laid out as a straddle the same day (rev 1's top/bottom split, rows directly opposite) and the board rerouted: DRC 0 unconnected, verify.py passes and now checks the straddle; renders regenerated. The row's distance from the edge, the pad size and the 1.6 mm thickness are estimates. Next: measure the parts and set `J1_ROW`, `J1_PAD` and `BOARD_T`. - 2026-11-19: Rev 2 draft case in hardware/sb-lora/rev2/case/: base and lid split at the board's top face, both printed flat with no supports, two M2 x 8 screws. The in-slot shape (rails, side latches, end tabs) is nkorth's public-domain Springboard slot cover (Thingiverse 4776750); past the slot opening a bump holds the XIAO and Wio, the LiPo below them, a 42 x 10 mm FPC antenna pocket and the USB-C opening. 54 x 77 x 17.4 mm assembled. Not printed. The connector's contact height in the slot (3.35 mm), J1's body, and the radio and LiPo heights are guesses to measure; the bump also needs checking against the Visor's IR window and stylus.