Robert A. James (Captain Bob) rajames
rajames commented on issue rajames/LithosAnanke#185 2026-11-20 14:46:23 +00:00
Build: map-64.f18 and map-32.f18 run the same sub-make twice under -j, and it races itself

Closed 2026-11-20, verified against the tree.

d783fd11: kernel/Makefile makes map-64.f18 and map-32.f18 one grouped target (&:, GNU make 4.3+), so their sub-make runs once. Seen to fail…

rajames closed issue rajames/LithosAnanke#182 2026-11-20 14:46:20 +00:00
Latent 32-bit block arithmetic in the stick's growth and the GPT reader
rajames commented on issue rajames/LithosAnanke#182 2026-11-20 14:46:17 +00:00
Latent 32-bit block arithmetic in the stick's growth and the GPT reader

Closed 2026-11-20, verified against the tree.

3c053f02: a compile-time check beside sk_v4_grow.c's (uint32_t)devblock * 4 (the tree's C99 negative-array pattern). An Artemis cap past 4 TiB…

rajames closed issue rajames/LithosAnanke#181 2026-11-20 14:46:16 +00:00
v4: comments that read as unfinished or say what the code no longer does
rajames commented on issue rajames/LithosAnanke#181 2026-11-20 14:46:14 +00:00
v4: comments that read as unfinished or say what the code no longer does

Closed 2026-11-20, verified against the tree.

90ea3728: heat.c's call_index() comment states the rule the code follows, instead of reading as an unfinished thought. stack.h no longer says there…

rajames closed issue rajames/LithosAnanke#180 2026-11-20 14:46:14 +00:00
gpt_read stops reading the entries at max, and leaves the rest unchecked
rajames commented on issue rajames/LithosAnanke#178 2026-11-20 14:46:13 +00:00
sk_v4_idle replaces the drive-slot array without copying it

Closed 2026-11-20, verified against the tree.

41f8648d: sk_v4_slots_get() (kernel/src/v4/sk_v4_slots.c) allocates a controller's slot array once, zeroed, at the size first asked for, and…

rajames closed issue rajames/LithosAnanke#178 2026-11-20 14:46:13 +00:00
sk_v4_idle replaces the drive-slot array without copying it
rajames commented on issue rajames/LithosAnanke#180 2026-11-20 14:46:13 +00:00
gpt_read stops reading the entries at max, and leaves the rest unchecked

Closed 2026-11-20, verified against the tree.

9700a7b4: gpt_read() range-checks every entry and counts every used one. A good table with more used entries than max is GPT_MANY, and the…

rajames commented on issue rajames/LithosAnanke#176 2026-11-20 14:45:47 +00:00
Zuse's key seeds are left in stack memory on the failure paths

Closed 2026-11-20, verified against the tree.

d7b2f6bf: the three functions share one buffer the identity module owns, g_seeds (the seed, then its read-back). sk_v4_drive_arrived, which every…

rajames closed issue rajames/LithosAnanke#176 2026-11-20 14:45:47 +00:00
Zuse's key seeds are left in stack memory on the failure paths
rajames closed issue rajames/LithosAnanke#170 2026-11-20 14:45:46 +00:00
USB: a device whose wTotalLength is 9 or less is asked for its configuration for ever
rajames commented on issue rajames/LithosAnanke#172 2026-11-20 14:45:46 +00:00
nightly.sh: the dict_hash agreement passes with fewer than three hashes

Closed 2026-11-20, verified against the tree.

c28cc019: scripts/nightly.sh dict-hash is a stage of its own, run by stick-high. It takes one hash from each of amd64, aarch64 and riscv64, all…

rajames closed issue rajames/LithosAnanke#172 2026-11-20 14:45:46 +00:00
nightly.sh: the dict_hash agreement passes with fewer than three hashes
rajames commented on issue rajames/LithosAnanke#170 2026-11-20 14:45:45 +00:00
USB: a device whose wTotalLength is 9 or less is asked for its configuration for ever

Closed 2026-11-20, verified against the tree.

a8824655: usb_config_request_length() asks for nothing more when wTotalLength is 9 or less. It says "xhci: a configuration of N bytes has no…

rajames closed issue rajames/LithosAnanke#169 2026-11-20 14:45:14 +00:00
v4_manage_serve takes its arguments without checking the stack's depth
rajames commented on issue rajames/LithosAnanke#169 2026-11-20 14:45:12 +00:00
v4_manage_serve takes its arguments without checking the stack's depth

Closed 2026-11-20, verified against the tree.

c642af54: v4/src/manage.c:108 refuses a request when the stack holds fewer values than it takes, or has no room for what it leaves (`n->ds.depth <…

rajames closed issue rajames/LithosAnanke#168 2026-11-20 14:45:10 +00:00
MINT-ZUSE says "nothing changed" after it has rewritten and zeroed the stick
rajames commented on issue rajames/LithosAnanke#168 2026-11-20 14:45:07 +00:00
MINT-ZUSE says "nothing changed" after it has rewritten and zeroed the stick

Closed 2026-11-20, verified against the tree.

54311be1 (with its protocol run, ff797d80):

  • kernel/src/v4/sk_v4_mint.c has erased(), "MINT-ZUSE: failed -- the stick was erased; Zuse not…
rajames commented on issue rajames/LithosAnanke#152 2026-11-20 14:43:50 +00:00
Gitea: a push to an open PR's branch can take 1.75 GB (patch test and diff statistics)

2026-11-20, as ruled ("write the practice, then close"):

  • The practice is written into .claude/CLAUDE.md, beside the developer protocol: while a PR is open, commit locally and push once. It is…