Units run in the wasm sandbox, so my_pid/parent_pid/context.server->request_count
read as sandbox stubs — the demo System Info counters were broken, and there was
no authoritative server-side request timing available to unit code (client-side
measurement cannot see queue/dispatch latency).
Add a request_perf() unit API backed by a new uce_host_request_perf hostcall.
The native worker answers it live, returning a DValue:
worker_pid, parent_pid, request_count,
accept_us = (time_start - time_init)*1e6 (entry -> dispatch wait),
running_us = (now - time_start)*1e6 (since dispatch, live),
total_us = (now - time_init)*1e6 (since the request entered UCE),
workspace_birth_us.
time_init is captured at request entry (handle_request, with a handle_complete
fallback); a RequestPerfSnapshot {pids, request_count, time_init, time_start} is
threaded from wasm_backend_serve through wasm_worker_serve onto the workspace,
and the hostcall computes the live deltas at call time. Wired like uce_host_units
(sized DValue hostcall): core_hostcalls.syms + sys.cpp/sys.h request_perf().
site/demo/index.uce System Info now uses request_perf() and shows the real worker
PID, an incrementing per-worker request count, and the timing counters.
Implemented via the pi agent (gpt-5.3-codex-spark); a review of the live numbers
caught accept_us mistakenly computed as (now - time_init) (== total_us), fixed to
the dispatch wait (time_start - time_init). Independently verified on the host:
System Info shows non-zero PIDs, incrementing count, accept_us ~50us << total_us
~2.4ms with accept+running==total; run_cli_tests --include-wasm-kill => 87 passed,
0 failed, 0 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follow-up to 560290c. Per-unit .wasm modules were still Cranelift-compiled per
worker on first use (~40-70ms each), so with 8-call worker recycling fresh
workers re-JIT every unit they touch.
Refactor the core cache logic into a shared WasmWorker helper:
- cached_wasm_path(p): maps <...>.wasm -> <...>.cwasm.
- load_or_compile_cached_module(engine, cached, wasm, bytes, err): deserialize_file
the .cwasm when it is newer than the .wasm; otherwise Module::compile + serialize,
written atomically (temp+rename); deserialize failure falls back to compile.
Both the core module load and unit_module() now go through this helper, so unit
artifacts get the same <unit>.uce.cwasm cache the core got.
Independently verified on the host: fresh-worker first-hit unit latency ~40-70ms
-> ~3-21ms; full suite wall-clock 20.5s -> 6.5s (and ~40s before any caching);
run_cli_tests --include-wasm-kill => 87 passed, 0 failed, 0 skipped.
Implemented via the pi agent (gpt-5.3-codex-spark sub-model).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Workers recycle every 8 requests (calls_until_termination=8). Each fresh worker
ran wasmtime::Module::compile() on the 6.8MB bin/wasm/core.wasm — a ~1.3s
Cranelift JIT — on its first request, so every ~8th request spiked to ~1.3s and
dominated suite wall-clock.
Cache the compiled artifact: on worker core-module load, if bin/wasm/core.cwasm
exists and is newer than core.wasm, load it via Module::deserialize_file() (mmap,
~ms); otherwise Module::compile() as before and atomically (temp+rename) write
the serialized artifact for the next worker. Deserialize failure / stale cache
falls back to a normal compile, so it is self-healing; rebuilding core.wasm
(newer mtime) invalidates the cache. Engine config (epoch_interruption,
signals_based_traps(false)) is unchanged, which the serialized format requires.
Implemented via the pi agent (delegated to a gpt-5.3-codex-spark sub-model).
Independently verified on the host: worker-startup request latency drops from
~1.3s to ~23ms (warm 12x /demo/hello.uce: all <=0.023s, no spikes); full suite
wall-clock ~40s -> 20.5s; run_cli_tests --include-wasm-kill => 87 passed, 0
failed, 0 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Post-deletion cleanup (units run only on wasm):
- types.h/compiler.cpp: drop the native-era SharedUnit fields so_name,
bin_file_name, and the opt_so_optional cache-mode plumbing (no native
optional .so path remains). The per-unit compile lock is re-keyed from
so_name+.lock to wasm_name+.lock (still per-unit).
- unit_info() and to_string(SharedUnit*) no longer expose .so artifact fields.
- backend.h: drop the stale "+ fallback-token gate" comment.
- Docs/comments corrected to wasm-only reality: README, tests/README,
site/doc C++ preprocessor + error_pages + unit_info pages, site/info intro,
site/demo/unit-browser artifact card; the Phase-5 native-vs-wasm benchmark
harness (tests/wasm_benchmark.py) reframed for the wasm-only backend.
Audit confirmed no live references remain to so_handle, load_shared_unit,
compiler_load_shared_unit, compiler_invoke*/_cli/_websocket/_serve_http,
COMPILE_SCRIPT/COMPILE_WASM_UNITS, or the native export-symbol constants;
request_ref_handler/dv_call_handler are kept (live wasm funcref casts).
Swept via the pi agent (delegated to a gpt-5.3-codex-spark sub-model);
independently re-verified on the host: run_cli_tests --include-wasm-kill =>
87 passed, 0 failed, 0 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Units now run exclusively on wasm; the native generated-C++ -> clang -> .so ->
dlopen path and its request-time fallback are removed.
- Dispatch (linux_fastcgi.cpp): the 4 handle_complete branches + the CLI-socket
path route every request through wasm (wasm_ready compiles cold/stale on
demand); a wasm-unavailable unit now yields a clean error page
(fail_wasm_unavailable / render_request_failure) instead of native execution.
compiler_invoke / _cli / _websocket / _serve_http deleted.
- compiler.cpp (-1274): removed the native .so compile (COMPILE_SCRIPT),
load_shared_unit, dlopen/dlsym/dlclose, compiler_load_shared_unit, and the
SharedUnit .so function-pointer fields (on_setup/on_render/on_component/
on_websocket/on_cli/on_once/on_init) in types.h/types.cpp. compile_shared_unit
now builds only the .wasm side-module; the .uce preprocessor/parser front-end
is kept (it emits the C++ the wasm compile consumes).
- unit_call()/component()/once/init now resolve across units through the wasm
host component resolver (uce_host_component_resolve) instead of native dlsym;
configured runtime error pages render through the wasm backend.
- Dropped the WASM_BACKEND_ENABLED feature flag and dead COMPILE_SCRIPT /
COMPILE_WASM_UNITS config; unit ABI freshness tied to UCE_UNIT_ABI_VERSION;
guard against serving a stale .wasm for a deleted source; retired the obsolete
W5 native-vs-wasm toggle script. Docs updated.
Implemented via the pi agent (with 3 delegated sub-reviews); independently
re-verified on the build host: run_cli_tests --include-wasm-kill => 87 passed,
0 failed, 0 skipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- WS: a dedicated broker process owns HTTP_PORT + every connection; it forwards
renders to the worker pool over uce.sock (non-blocking) and applies ws_*
command batches flushed back at workspace teardown. Removes the now-dead
per-worker websocket executor (-509 lines).
- Dispatch: unify CLI / WebSocket / serve_http / page render through one
serve_via_wasm(entry_unit, handler) path; handler string -> __uce_<handler>
export symbol.
- W7d: rewrite zip.uce to the membrane return-value error contract (no C++
try/catch), error-reporting.uce to genuine wasm traps instead of throw, and
sharedunit.uce to unit_info(); empty the native-only token gate.
- Membrane: wire ls / mkdir / file_mtime through new uce_host_file_list /
uce_host_file_mkdir / uce_host_file_mtime hostcalls (resolve_guest_file gains
directory support). Fixes /doc/index.uce listing nothing; adds a regression
assertion that the index enumerates items.
- Docs: add docs/wasm-runtime-architecture.md; record the W7e staged native-
deletion plan in WASM-PROPOSAL.md.
Verified: scripts/run_cli_tests.sh --include-wasm-kill -> 87 passed, 0 failed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
