Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
1d43beeabf
hardware: SB-LoRa rev 2 J1 as a straddle for the MCS-068SMA1: pins 1-34 on F.Cu, 35-68 on B.Cu directly opposite at 1.27 mm (rev 1's split); row position, pad size and board thickness are estimates until a part is measured. Rerouted: DRC 0 unconnected, verify.py checks the straddle and passes; renders regenerated
7b4e23ad64
docs: progress: SB-LoRa rev 2 milestones (planned, laid out, J1 chosen and ordered), the cancelled rev 1 J1, the release workflow, and the J1 supply and board-thickness questions
fdacc2daa8
docs: SB-LoRa rev 2 connector: the MCS-068SMA1 straddle parts are ordered (2026-11-19) and J1 is designed around them; no stock of the single-row MCS-068SSA1 turned up
f41ddf43bd
docs: SB-LoRa rev 2 connector: the MCS-068SMA1 is Robinson Nugent's MEMPAK straddle receptacle (.050" pitch, no tail offset) per the catalogue overview page; the row gap still needs a sample, and the single-row MCS-068SSA1 is noted as a single-sided alternative
ef1d245c73
docs: SB-LoRa rev 2 connector: 3M MCS-068SMA1-TR30 (female straddle receptacle, eBay surplus, rows not staggered) replaces the AliExpress part as the J1 candidate, which is dropped for $39 shipping and a 2+ week wait; the J1 pattern becomes opposed F.Cu/B.Cu rows and the board thickness follows the connector's row gap
fff912fb4d
docs: SB-LoRa rev 2 connector: the J1 listing photo shows an AMP-marked part with no part number but it's sold as Jiansu BST, so the land pattern comes from the parts that arrive and they're bought in one lot
e71f979dce
docs: SB-LoRa rev 2 connector: the AliExpress 68-pin female PCMCIA part is the J1 candidate (female side, both rows of tails flat on top per the listing photos); its land pattern waits on a measured sample, and the Hirose IC1K listing looks like a host-side socket
fec5a03358
palm-client: release workflow: a v* tag builds and tests the app and publishes visormesh-<tag>.prc with SHA256SUMS as a release; the toolchain install moves to tools/install_toolchain.sh so CI and the Dockerfile share the pinned versions
00feaffca5
hardware: SB-LoRa rev 2 footprints for the XIAO (flat or by its header, bottom BAT/SWD pads through solder-filled holes), the Wio (14 holes) and the LiPo pads, generated from Seeed's own KiCad files
bbfb334431
hardware: SB-LoRa rev 2 LCSC numbers for the flash (C914935), OR gate (C91874), inverters (C105251) and dual translator (C15741)
440845684c
hardware: SB-LoRa rev 2 flash pinout checked against the S29AL016J datasheet; it reads as an Am29LV160BB to Handspring's flash driver
de88c7a7a1
hardware: SB-LoRa rev 2 netlist draft: S29AL016J flash on CS0, SC16C650B UART on CS1 with OR-gated strobes (the slot can assert OE*/WE* with CS1*, the UART wants 10 ns after CS), XIAO and Wio side by side, battery and SWD pads. Plan: the flash's installer copies the app to RAM, since the transport's interrupt handler can't live on a removable module; battery sizing estimate
0136ed96ab
docs: layout fixes and New Key verified on the Neo
52f3f09a13
docs: AGENTS.md covers the try-once check and make replay
4b05bf25d5
palm-client: Add Channel's New Key (private channels) fills in a random key from a pool fed by what can't be predicted: the exact tick and position of every pen and key event, and the signal readings of every packet the radio reports. Samples are folded into a SHA-256 pool, a key is a hash of it, and the pool moves on after each; it's kept between launches in an unsaved preference (never in a backup). Until 256 samples have come in, New Key asks for a few seconds of stylus movement. Host-tested in tests/test_entropy.c
2c32033292
palm-client: check_ui.py fails the build when two drawn form objects overlap (a label running into its value or a button); it reads each object's usable bit so popup lists shown by the code don't count
18c55ef8aa
palm-client: three labels no longer run into what follows them: Node's 'Firmware:' is 'Version:', Preferences' Check when closed trigger moved right (its choices are '5 min' to '1 hour' so the longest still fits), and Flood Scopes' default moved right
bb2fc014ab
palm-client: the module can't keep waking a Visor on failing batteries. A slot power handler (SbPwr, in the copied block) stops the chip interrupting on a low-voltage shutdown, as the Springboard guide asks of modules without LOWBAT*; since the Neo's OS never reports that reason, the wake's card event handler and BgArmWake also check the battery and turn the wake off at the OS's warning level. Host-tested; untested on hardware
6846d4d79b
docs: Slot restore with an empty slot verified on the Neo; wake and bring-up plan noted
db46a0806f
hardware: SB-LoRa bring-up plan (BRINGUP.md): bare-board and bench-power checks, what each SB-LoRa Test line means and what a failure points to (including the A0 mapping and the read values), moving the XIAO stack over, the Slot link, the message wake and standby current
e86b14e440
palm-client: a saved Slot link is restored at launch while the OS reports a module (checked without touching the bus), so the closed-app check and its wake use the module; with the slot empty it stays off, as before, so a launch can't reset the Visor
f53f851b0e
palm-client: with the Slot link, the radio's 'message waiting' push wakes the Visor. The slot interrupt handler (SbIrq) spots the frame as it drains the FIFO and, as the Springboard guide describes, holds IRQ* (THR empty) to have the system woken, then keeps it awake 10 s for the app to sync. With Check when closed on, the app leaves a copy of the handler in system-owned memory as it closes (SbWakeArm), with the module asleep between frames; it posts a card event whose handler sets the app's alarm for now, so the closed-app check runs at once. check_sections.py checks that the copied block (SbCoreStart..SbCoreEnd) is whole and refers only to itself. The Slot link no longer reconnects after sleep. Host-tested on the simulated module; untested on hardware
1354b6531a
docs: Trace result, command retry, empty-slot SB-LoRa Test and Add Contact note verified on the Neo
3ca184099c
docs: contact login status verified on the Neo
6056339ad8
palm-client: the contact screen's login status is 'Logging in over the mesh...' (it ran off the edge) and its status line is cut rather than running off; Serial Ports' lines too. check_ui.py now measures literals passed to a screen's own one-line helpers, and fails on a new helper that draws its text until its width is listed
9ae5decc84
docs: Node clock and Radio Stats layout verified on the Neo
5aa68d46ec
docs: AGENTS.md describes check_source.py and the screen line tests
14fc4ffe0c
docs: progress: connection state in conversations, reconnecting in the background, the USB send block
52c4101796
palm-client: over USB, don't send while CTS is off: it's off while nothing on the host reads (bridge stopped), and then SrmSend blocks 10 s whatever the CTS timeout (measured on the Neo: that was the slow reconnect). Reconnect steps skip the send; other sends wait up to 1 s for the host. Drops the CTS-timeout setting, which the USB driver ignores
453559cd5c
palm-client: a reconnect in steps is lighter on the app: reads that can't wait check what's buffered first (SrmReceiveCheck), steps 4 times a second instead of 10, its DEVICE_QUERY send waits 100 ms instead of 250
b5ce1098b3
palm-client: reconnects after a wake or a lost link run in steps on idle events (open, attach wait, DEVICE_QUERY with the reply checked without waiting), so the app stays usable while it looks for the radio; titles show "connecting"; Options > Connect and a tap on "offline" still connect in one go
f9aa7a1e93
palm-client: serial sends wait 250 ms at a time (8 tries, 2 s) instead of the 5 s CTS default, so a connect with nothing reading at the other end (bridge stopped) can't freeze the app for over a minute