Repository navigation
fix(gust): restore gust:os syscall-seam dissolve path (time-provider resolution() drift from #202) - #213
Merged
Conversation
…all seam (REQ-OS-SYSCALL-001) PR #202 (gust:os timer-sleep) added `resolution: func() -> u64` to the `time` interface in wit-os/gust-os.wit but never updated the `time-provider` crate's `impl Guest for P`, so it silently broke `cargo build` for time-provider (E0046, missing trait item) — and with it BOTH tl-node (build-os-tl.sh) and ts-node (build-os-ts.sh), since both import `time`. This regression shipped on main because gust_timer_probe (#202's own demo) drives the executor natively and never exercises the time-provider dissolve path, so nothing caught it. Adds resolution() returning a documented placeholder tick rate (1 MHz) — TIM2_CNT here is a raw mocked counter with no real PSC/ARR clock-tree config crossing this seam, so there is no genuine Hz to report yet. With the fix, re-verified end to end on the CURRENTLY INSTALLED toolchain (synth 0.49.0, loom 1.2.0, meld 0.41.3 — all post meld#326/synth#757 fixes): - build-os-tl.sh: os-tl-cm3.o text=2206 data=52 bss=3096 (5354B/8192) - gust_os_tl_probe (qemu): "gust-os-tl-probe OK: log==\"gust:os up\\n\", run()=0xA" — buffer-carrying log.line crosses the fused-memory boundary UNCORRUPTED, exit 0. - build-os-ts.sh: os-ts-cm3.o text=1612 data=40 bss=2584 (4236B/8192) - gust_os_ts_probe (qemu): "gust-os-ts-probe OK: poll==1", exit 0. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011QG86sovTbfnPNY9SfhSmo
Contributor
Author
|
Re-triggering CI: the self-hosted runner went idle and #213's jobs (queued since 06:10) were zombied — reopening to spawn fresh runs against the live runner. |
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
avrabe
added a commit
that referenced
this pull request
Jul 23, 2026
…32 (gale#214) (#220) Closes the process gap that let #202 ship green: it added time.resolution() to the gust:os WIT without the time-provider Guest impl (E0046), silently breaking the gust_os_tl/ts dissolve until #213 caught it by hand. No workflow built the provider crates, so the compile error never fired in CI. This gate builds every gust:os provider/app crate for wasm32 (the dissolve target) — a WIT method that outruns its Guest impl now fails RED. Cargo-only; the wit-bindgen-generated Guest trait makes the compile the oracle. Red-tested: injecting a method into the time interface without the impl reproduces E0046 'not all trait items implemented'; the gate on main is green (8/8 crates). Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Workstream B of the v0.5.0 release — restores the
gust:ossyscall-seam dissolve path and confirms it works end-to-end on the now-fixed toolchain.The bug (silently shipped)
PR #202 (gust:os timer-sleep) added
resolution: func() -> u64to thetimeinterface inwit-os/gust-os.wit, but never updated thetime-providercrate'simpl Guest for P. That's a missing trait item (E0046):cargo buildfortime-providerwas broken — and with it both the tl-node and ts-node dissolves (both importtime). It shipped unnoticed because #202's own demonstrator drives the executor natively and never exercises thetime-providerdissolve path.Fix
time-provider/src/lib.rs: addfn resolution() -> u64 { RESOLUTION_HZ }(RESOLUTION_HZ = 1_000_000), documented as a placeholder — TIM2_CNT here is a mocked free-running counter with no real PSC/ARR clock-tree config crossing this seam, so there is no genuine Hz to report yet. Regenerated the affected dissolved cores (os-tl-cm3.o,os-ts-cm3.o) with the current toolchain.Verified (the decisive test — the seam was blocked for months on meld#326 + synth#757, both now fixed)
On synth 0.49.0 / loom 1.2.0 / meld 0.41.3:
build-os-tl.sh→os-tl-cm3.otext 2206 / bss 3096 (5354 B / 8192 budget)gust_os_tl_probe(qemu) →OK: log=="gust:os up\n", run()=0xA— the buffer-carryinglog.linecapability crosses the fused-memory boundary uncorrupted.gust_os_ts_probe(spawn) →OK: poll==1.Follow-ups (not in this PR)
VER-OS-SYSCALL-001evidence link + a Renode content-gate for the seam.resolution()becomes a real silicon-derived value once a configured-PSC timer is wired into the seam.🤖 Generated with Claude Code