2026-10-30 - 2026-11-06
Overview
1 Release published by 1 user
Published
archive/starforth-2026-09-17
1 Pull request merged by 1 user
Merged
#30 Merge StarForth-v4.0.0 into master
1 Pull request proposed by 1 user
Proposed
#110 The record of 2026-11-06, and issue and PR templates
68 Issues closed from 1 user
Closed
#111 [ruling] The public demo's console: a demo identity, or an emergency-console build?
Closed
#107 One repository: StarForth archived as a tag and deleted; this one renamed LithosAnanke
Closed
#106 The acceptance gate on master
Closed
#105 CI on the runner: every job failed -- ubuntu-latest is Debian 11
Closed
#104 The nightly snapshot and the release, to Nexus
Closed
#103 make -C docs/formal doxygen fails: newunicodechar, ASCII character requested
Closed
#102 make init: install the toolchain, syft and Isabelle2025-2
Closed
#101 1,003 logs and DoE CSVs committed as plain files, outside LFS
Closed
#99 A USB keyboard; and a PC without PS/2 hung at boot
Closed
#100 Review of the v4 golden model; three bugs fixed (line input, refused stores)
Closed
#98 The USB controller found as real hardware has it; every controller, not the first
Closed
#97 The Pi 5's native path: MMU and caches before any C, bounded mailbox, framebuffer from the devicetree
Closed
#91 make write_stick: the PC stick onto a USB stick, checked and read back
Closed
#90 The amd64 image page-faults when a PC loads it above 2 GiB
Closed
#89 Stick increment 4: Artemis and Zuse from one disk
Closed
#88 Stick increment 3: the images -- the stick and the data stick, Artemis signed at build
Closed
#87 Stick increment 2: the GPT reader, and Zuse from a GPT identity partition
Closed
#86 Stick increment 1: the keyboard bridge shared, and v4's screen
Closed
#82 The identity gate: no identity drive, headless; Zuse's drive, a console
Closed
#81 MESH step 7: a second unit, PAUSE, RESUME, KILL
Closed
#80 MACHINE steps 1-2: the ROM assembles the system at boot; the mesh ROM
Closed
#79 v4 is the default build; both systems on every board image
Closed
#78 MESH step 6: storage -- every node asks the kernel for its blocks (6, 6b-6e)
Closed
#77 MESH steps 1-5: ports and the fabric, the nucleus as a capsule, messages, finding the way, birth
Closed
#76 ENGINE steps 1-2: interpret this text, ask the kernel
Closed
#75 The golden model: one node, all 32 opcodes, heat, POST at 32 and 64 bits
Closed
#73 LOG-FLUSH: a VM's log records reach Artemis's log ring
Closed
#74 StarForth v4: the justification and the decomposition of every v3 word
Closed
#72 A message could run twice; one delivery path, a VM takes its own messages on its own turn
Closed
#71 Challenge-response against a cloned identity drive
Closed
#69 The elevation entrypoint let any identity run FORTH as Zuse
Closed
#70 Refuse a same-VM, cross-device raw block copy
Closed
#67 CRLF submits a line twice; zuse_root_pubkey_known never set after a same-session genesis
Closed
#68 Reshuffle phases 4-5: remove the Hermes VM; Isabelle pass, docs, SBOM, tag v2.1.0
Closed
#65 Reshuffle phase 3: cut every FORTH-owned message type over to kernel-Hermes
Closed
#64 Reshuffle phase 2: the kernel-Hermes allocator and its audit
Closed
#66 SkHermesMessage payload aliasing
Closed
#63 Reshuffle phase 1: Hestia
Closed
#61 The Tripod/kernel reshuffle: design (Hermes into the kernel; Hera, Artemis, Hestia)
Closed
#62 Reshuffle phase 0: preparation, no behaviour change
Closed
#59 THRU swallows lines
Closed
#57 Human and unattended identities; USE fail-closed halt
Closed
#56 std79 DoE rerun after stage 4: 81 of 81
Closed
#55 Preemptive context switching between VMs (stages 0-4)
Closed
#54 Per-VM log persistence, with source attribution
Closed
#53 Ephemeral-birth workload design; BIRTH/CAPSULE-BIRTH extended symmetrically
Closed
#52 mkcapsule --resolve: block-number collisions are silent persistent-disk corruption
Closed
#50 riscv64-only stall in the 9-identity campaign: an O(N) idle pump
Closed
#49 The turn-attractor rebuilt on real messaging
Closed
#48 Hera can't message -- fixed by symmetry, not special-cased
Closed
#47 VM-HEAT: groundwork for a compudynamic turn-attractor
Closed
#46 The fleet conservation invariant K, verified continuously
Closed
#45 stadium_grant_quota(): the donor floor
Closed
#44 aarch64 Stadium/COOL scaling stall
Closed
#43 std79 DoE: 3 architectures x 9 identities x 3 replicates, randomised full factorial
Closed
#41 Heap corruption under repeated WIREBIND attach/detach
Closed
#40 Cross-ISA standard-dictionary exerciser campaign -- two bugs found live
Closed
#39 FORTH-79/83 identity lockdown, and Hera to Artemis storage-attach messaging
Closed
#38 xHCI concurrent enumeration -- the 4th-device failure was QEMU's port layout
Closed
#37 Four bugs found minting and verifying the 8 identity thumbdrives
Closed
#36 Headless until login, and the KILL / g_repl_active_vm dangling pointer
Closed
#35 xHCI/BOT driver: one device at a time must become many
Closed
#34 aarch64 BYE traps: PSCI SYSTEM_RESET issued by smc with no secure monitor
Closed
#33 Artemis Stadium migration (item 4.6) fails every trial: quota granted after IDENTITY runs
Closed
#32 Version correction: the v2.0.1 bump and v2.0.0 tag were premature
Closed
#31 Merge v2.0.1 into master and verify the build and behaviour are the same
Closed
#19 Merge StarForth-v4.0.0 into master
Closed
#18 master's build.yml: old tree layout, and a second 02:00 nightly
112 Issues created by 1 user
Opened
#6 Word-level ACL on the v4 path: the countdown and recheck at call (ENGINE.md step 3)
Opened
#4 v4 parity with v3: 317 words against v3's 526, and v3's POST not yet run on v4
Opened
#5 Measure K=1.0 on the golden model, at 32- and 64-bit cells, on all three host ISAs
Opened
#10 Kernel words for the v4 node: v3's C words served from a node (ENGINE.md step 4)
Opened
#11 Isabelle/HOL proofs for v4: opcode semantics, and the stack and fault model
Opened
#12 A byte address given to ! -- PAD 42 OVER ! (D-1)
Opened
#13 Blocks that survive a power-off on bare metal
Opened
#14 Kernel block requests push onto the node's stack without the overflow check (D-16)
Opened
#8 Verilator co-simulation of one node, and a per-instruction state trace from the golden model
Opened
#15 Fuzz the text assembler, the ROM's assembler and the line input
Opened
#7 Hosted mesh running the Tripod roles as nodes (ENGINE.md steps 6-7)
Opened
#9 Rule D-4 (the memory map) and D-6 (heat structures in hardware)
Opened
#16 Run make -C v4 sanitize and hosted-check in CI
Opened
#17 First nightly run: confirm the Nexus publish
Opened
#20 Check build parameters at compile time: V4_PORTS above 16 overflows the port-status bits
Opened
#24 v4/README.md: stale status lines
Opened
#18 master's build.yml: old tree layout, and a second 02:00 nightly
Opened
#19 Merge StarForth-v4.0.0 into master
Opened
#21 An error raised to an unattached error register is lost
Opened
#22 One implementation of the wire operations (boot.c and fabric.c)
Opened
#23 Strict warnings on boot.c, post.c and blocks.c
Opened
#29 A Raspberry Pi Imager repository for the nightly images
Opened
#25 LaTeX build output committed in docs/
Opened
#28 Raspberry Pi 5: boot the nightly image on the board
Opened
#26 Raspberry Pi 4 port (BCM2711)
Opened
#27 Milk-V Mars: first boot on the board
Opened
#31 Merge v2.0.1 into master and verify the build and behaviour are the same
Opened
#32 Version correction: the v2.0.1 bump and v2.0.0 tag were premature
Opened
#33 Artemis Stadium migration (item 4.6) fails every trial: quota granted after IDENTITY runs
Opened
#34 aarch64 BYE traps: PSCI SYSTEM_RESET issued by smc with no secure monitor
Opened
#35 xHCI/BOT driver: one device at a time must become many
Opened
#36 Headless until login, and the KILL / g_repl_active_vm dangling pointer
Opened
#37 Four bugs found minting and verifying the 8 identity thumbdrives
Opened
#38 xHCI concurrent enumeration -- the 4th-device failure was QEMU's port layout
Opened
#39 FORTH-79/83 identity lockdown, and Hera to Artemis storage-attach messaging
Opened
#40 Cross-ISA standard-dictionary exerciser campaign -- two bugs found live
Opened
#41 Heap corruption under repeated WIREBIND attach/detach
Opened
#42 Several identity drives attached at once: only some are detected
Opened
#43 std79 DoE: 3 architectures x 9 identities x 3 replicates, randomised full factorial
Opened
#44 aarch64 Stadium/COOL scaling stall
Opened
#46 The fleet conservation invariant K, verified continuously
Opened
#45 stadium_grant_quota(): the donor floor
Opened
#47 VM-HEAT: groundwork for a compudynamic turn-attractor
Opened
#48 Hera can't message -- fixed by symmetry, not special-cased
Opened
#49 The turn-attractor rebuilt on real messaging
Opened
#50 riscv64-only stall in the 9-identity campaign: an O(N) idle pump
Opened
#51 Raise the DoE's N-REPS above 3
Opened
#52 mkcapsule --resolve: block-number collisions are silent persistent-disk corruption
Opened
#53 Ephemeral-birth workload design; BIRTH/CAPSULE-BIRTH extended symmetrically
Opened
#54 Per-VM log persistence, with source attribution
Opened
#55 Preemptive context switching between VMs (stages 0-4)
Opened
#56 std79 DoE rerun after stage 4: 81 of 81
Opened
#57 Human and unattended identities; USE fail-closed halt
Opened
#58 Multiuser/multitasking DoE: put the OS through its paces
Opened
#59 THRU swallows lines
Opened
#61 The Tripod/kernel reshuffle: design (Hermes into the kernel; Hera, Artemis, Hestia)
Opened
#60 Block-backed DoE persistence, for bare metal with no host to scrape the serial log
Opened
#62 Reshuffle phase 0: preparation, no behaviour change
Opened
#64 Reshuffle phase 2: the kernel-Hermes allocator and its audit
Opened
#63 Reshuffle phase 1: Hestia
Opened
#66 SkHermesMessage payload aliasing
Opened
#65 Reshuffle phase 3: cut every FORTH-owned message type over to kernel-Hermes
Opened
#67 CRLF submits a line twice; zuse_root_pubkey_known never set after a same-session genesis
Opened
#68 Reshuffle phases 4-5: remove the Hermes VM; Isabelle pass, docs, SBOM, tag v2.1.0
Opened
#70 Refuse a same-VM, cross-device raw block copy
Opened
#69 The elevation entrypoint let any identity run FORTH as Zuse
Opened
#71 Challenge-response against a cloned identity drive
Opened
#72 A message could run twice; one delivery path, a VM takes its own messages on its own turn
Opened
#73 LOG-FLUSH: a VM's log records reach Artemis's log ring
Opened
#74 StarForth v4: the justification and the decomposition of every v3 word
Opened
#75 The golden model: one node, all 32 opcodes, heat, POST at 32 and 64 bits
Opened
#76 ENGINE steps 1-2: interpret this text, ask the kernel
Opened
#77 MESH steps 1-5: ports and the fabric, the nucleus as a capsule, messages, finding the way, birth
Opened
#78 MESH step 6: storage -- every node asks the kernel for its blocks (6, 6b-6e)
Opened
#79 v4 is the default build; both systems on every board image
Opened
#80 MACHINE steps 1-2: the ROM assembles the system at boot; the mesh ROM
Opened
#81 MESH step 7: a second unit, PAUSE, RESUME, KILL
Opened
#82 The identity gate: no identity drive, headless; Zuse's drive, a console
Opened
#84 KEY, EXPECT and QUERY read messages, not the console registers (MESH step 3)
Opened
#83 MACHINE steps 3-6: init.4th after POST; Hera streams the ROM; FORTH-79 out of the .f18; v3 out of the build
Opened
#85 MESH steps 8-9: the products are the five nodes, hosted and then bare metal
Opened
#86 Stick increment 1: the keyboard bridge shared, and v4's screen
Opened
#87 Stick increment 2: the GPT reader, and Zuse from a GPT identity partition
Opened
#88 Stick increment 3: the images -- the stick and the data stick, Artemis signed at build
Opened
#89 Stick increment 4: Artemis and Zuse from one disk
Opened
#90 The amd64 image page-faults when a PC loads it above 2 GiB
Opened
#91 make write_stick: the PC stick onto a USB stick, checked and read back
Opened
#93 Stick increment 6: pull and replug, and the flush
Opened
#92 Stick increments 5 and 5a: genesis on the stick, growth, the proxy Zuse and the real Zuse
Opened
#94 Stick increment 7: the splash and progress bar
Opened
#95 Stick increment 8: x86_64 acceptance moves to the stick boot
Opened
#97 The Pi 5's native path: MMU and caches before any C, bounded mailbox, framebuffer from the devicetree
Opened
#96 SER5: boot the new PC stick
Opened
#98 The USB controller found as real hardware has it; every controller, not the first
Opened
#99 A USB keyboard; and a PC without PS/2 hung at boot
Opened
#100 Review of the v4 golden model; three bugs fixed (line input, refused stores)
Opened
#101 1,003 logs and DoE CSVs committed as plain files, outside LFS
Opened
#103 make -C docs/formal doxygen fails: newunicodechar, ASCII character requested
Opened
#102 make init: install the toolchain, syft and Isabelle2025-2
Opened
#104 The nightly snapshot and the release, to Nexus
Opened
#105 CI on the runner: every job failed -- ubuntu-latest is Debian 11
Opened
#106 The acceptance gate on master
Opened
#107 One repository: StarForth archived as a tag and deleted; this one renamed LithosAnanke
Opened
#108 Gitea issue forms and a PR template in the project's own form
Opened
#109 Revoke the access token claude-session-20261106
Opened
#114 Demo: the web terminal (ttyd) as a service
Opened
#113 Demo: instant start -- sessions resume from a saved, booted machine
Opened
#112 Demo: the session launcher -- one throwaway QEMU per visitor
Opened
#116 Demo: capacity -- how many visitors at once
Opened
#111 [ruling] The public demo's console: a demo identity, or an emergency-console build?
Opened
#115 Demo: Apache on the .com -- TLS and the websocket proxy
Opened
#117 Demo: the XWiki "Try it" page, with the console embedded