The manual data is invariant — derived only from the embedded manual — so
each View re-parsed it needlessly. It's now computed once and shared by all
Views and the LLM.
The intro and global-settings achievements recorded state even while
achievements were disabled, so the intro toast could be consumed before the
user ever enabled the feature — and then never appear when they later opted
in. The intro now fires at the enable points, so it shows the first time the
feature is enabled.
Achieve() now propagates a category unlock to the category's remaining
items, matching the constructor's load-time behavior. This lets an
achievement complete in the session where its category first unlocks.
The live per-frame scan visits only the set of locks with nonzero wait
counts, updated at each lock event: every wait-for edge originates at a
waiter, so uncontented locks cannot contribute anything. File loads replay
events without touching the set and are covered by the one-shot full-map
scan at end of load.
The scan consumes the final Waiting/SharedHolding flags of each lock
map: a thread blocked in a lock acquisition emits no Obtain and no other
event, so an unclosed wait at the last known state is a permanent wait
per the event contract. A failed try_lock never emits a wait, so it
cannot fabricate an edge. Locks recorded by clients with inverted
release ordering are skipped entirely, as their holder attribution is
ambiguous.
A thread that holds a shared lock and waits for exclusive cannot
progress while its own shared hold persists; this is reported as a
size-1 cycle.
The differences against the previous renderer are deliberate: the
drawn state is canonical from event 0, HasBlockingLock clears when the
blockers drain, same-state neighbors merge, a waiting thread shows as
holding only on its true acquire, a shared holder waiting for exclusive
draws as blocking, the Blocked and Sharing lists contain only genuinely
blocked roles, and the tooltip holder count follows the cursor position
in the holder-change log.
Capture has serialized lock events since d6f32a083 and 3e3aa80fa, so
the out-of-order reconstruction machinery has nothing left to repair.
Mark now targets the recorded Wait/Obtain event index; the old reverse
scan could walk off the start of the timeline for a thread that never
waited.
InsertLockEvent wraps the uint8 slot index past the 64-entry
per-lock table, misattributing the 65th distinct thread to slot
0. Set the failure instead.
With unlock -> notify order the thread can be preempted by other thread,
for example one that wants to lock the very same lock. This would produce
incorrect ordering of lock release and acquire events.
The deploy job SFTPs the emscripten build using repo secrets
(SERVER/USERNAME/PRIVATE_KEY/REMOTE_PATH) that only exist upstream.
On a fork these are empty, so every push to a fork's master fails
with "remote_path is empty" even though the actual build succeeded.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVBiYgtsB6okBueTwp5swa
