The native platform block is closed at 5fcc111 (GitHub run 30). Update the
work-package rows and open blockers: the remaining release gates are the remote
G3 artifact and a MEASURED G4 report, not Windows/macOS. Keep the historical
sections and their older SHAs intact.
GitHub run 30 on 5fcc111 is green for Linux, windows-latest and macos-latest
(the native failures were Windows PATHEXT/temp/rooted-path issues, macOS test
temp canonicalization and a Windows 1 MiB main-thread overflow). Woodpecker
#25 is green on the same SHA, and the disposable-VM G3 lane PASSes 7/7 plus the
strict JSON/inert-@knowledge contracts.
The release decision stays DO_NOT_RELEASE, now only for the remote G3 artifact
with matching tested_sha and the MEASURED G4 report; the native platform block
is closed. Document that the adversarial suite is G3-opt-in so ordinary
cargo test --workspace cannot turn a hosted namespace-capable host into a G0
failure.
A drive-less rooted path such as /tmp/x or \tmp\x is not Path::is_absolute()
on Windows, but joining it still escapes the project root. Validate with
has_root() so every rooted dependency path is rejected on every platform. The
existing rejection test now also covers the Windows root-relative form.
fs::canonicalize on Windows returns a \\?\C:\... path. Node's module loader
lstats the working/script path and rejects the verbatim form, so all Node
foreign-kernel tests failed on windows-latest. Use the native temp directory on
Windows and retain canonicalization on Unix for symlink safety; widen the CI
annotator so any remaining native failure shows a full assertion block.
Windows gives the main thread 1 MiB, unlike Linux/macOS at 8 MiB. The bundled
regex self-test overflowed that stack even though the same document passed on
Linux, so run the CLI dispatch on an explicit 8 MiB thread on every platform.
The macOS package test compared a raw temp path with a canonicalized one; on
macOS /var is a symlink to /private/var. Compare canonical forms on both sides.
The adversarial security matrix is G3-only: ordinary cargo test --workspace on
a namespace-capable but non-disposable hosted runner turned G0 red. Make it
opt-in (MARQDO_REQUIRE_STRICT_SECURITY=1 for the G3 lane, or
MARQDO_RUN_STRICT_SECURITY=1 for a developer); the lane still requires the
mechanism and fails on a missing one.
macOS arm64 aborted in nesting_depth_limit_is_enforced because the test-harness
thread stack is smaller than a production main thread. Run that contract test
on an explicit 8 MiB stack; the 128-level parser bound itself is unchanged.
The stdlib process product test hard-coded /bin/echo, which does not exist on
Windows. Use the test CLI itself as the granted program: the fixture reads
MARQDO_TEST_PROCESS_PROGRAM from the granted env, runs it with --version, and
the harness compares against the same invocation. process.run now forwards
SYSTEMROOT/WINDIR on Windows (process creation/DLL loading, not a capability),
and the bundled stdlib digest is regenerated for the changed test fixture.
The built-in Python provider scanned PATH for the extension-less token
python3; a Windows PATH contains python3.exe, so a working interpreter was
reported as kernel.unavailable and the semantic golden suite failed on
windows-latest. Expand candidates with PATHEXT (conservative default) and use
the same platform-aware resolver in the conformance test's availability probe.
Also make native failures diagnosable: strip ANSI before matching and emit the
panic/build context as public check annotations, and run Linux tests through
the same annotator. This fixes the Windows native lane and gives the macOS
failure the context needed to diagnose it.
The native matrix only annotated the first matching line, which omitted the
fixture name and expected/actual values. Annotate the panic block and build
error context instead so the failing golden is identifiable from the public
check-run annotations.
The Windows/macOS jobs fail inside cargo test --workspace, but public check
annotations only show the step exit. Tee the output and emit the failing test,
panic or build-error line as a check annotation so the native failure is
diagnosable without repository-admin log access. Behaviour and exit status are
unchanged.
Record the 512b9f5 local disposable-VM run (7/7 plus strict JSON and inert
@knowledge, evidence.json/archive/log digests) and Woodpecker pipeline #16.
The lane now fails if strict-evidence.json cannot be written, so the green run
also proves the evidence path. Fix the cross-platform JSON smoke to tolerate
the expected non-zero strict refusal exit before parsing.
Tested code SHA: 512b9f5. Release decision remains DO_NOT_RELEASE for the
missing remote Windows/macOS native PASS, remote G3 artifact and MEASURED G4.
A green lane without a strict-evidence.json must not count as a gate. The EXIT
trap now turns a generator failure into a lane failure. The self-hosted
workflow exports the archive path in the run step (env written by GITHUB_ENV is
not guaranteed in the env expression context), records the pinned image digest
as an input, and verifies schema/outcome/tested_sha/non-root before uploading.
Also define SOURCE_SHA before the VM overlay commands.
Record the six-step RFC-CI-0001 series, the Woodpecker pipelines, the Windows
cross-target check, the local disposable-VM G-C PASS at 0d7e11b (7/7 plus the
strict JSON and inert-@knowledge contracts), and the correctness-verified
performance SMOKE run. Keep the release decision at DO_NOT_RELEASE because the
same SHA still lacks a remote Windows/macOS native PASS, a remote G3 artifact
and a MEASURED G4 report.
Include the RFC itself and link it from docs/ci.md. This is the documentation
follow-up to step 6; the tested code SHA is 0d7e11b.
Add tools/ci/release_evidence.py (marqdo.release-gates.v1) plus unit tests:
missing/skipped/FAIL/UNVERIFIED and tested_sha mismatches stay
DO_NOT_RELEASE, performance may be PASS/WARN, unknown gate names are rejected.
The GitHub CI now has a release-evidence job that maps the Linux/schema and
native matrix results onto the manifest and uploads it, and both CI systems
run the evidence-tool tests.
Clear the contradictions the RFC calls out: remove the stale CURRENT.md rows
that claimed DNS/TLS/UDP and fs metadata were absent while the same table said
they were implemented; correct docs/ci.md to state that the GitHub-hosted
Windows/macOS matrix exists (Woodpecker has no such agent); keep the old
convergence plan historical; require the evidence manifest in releasing.md;
and add a docs-consistency guard for these claims.
This is step 6 of the CI/security RFC (B05/B06/B07).
Replace the three-run baseline profile with tools/perf/runner.py: one fixed
release binary, per-child rusage via posix_spawn/wait4, 5 warmups and 20
samples by default, SHA-derived interleaved order, p50/p95/MAD, CPU user/sys
and peak RSS. Every warmup output is normalized and compared with a committed
wire digest before timing; a mismatch is CORRECTNESS_FAIL, never a fast
sample. Short or unapproved-host runs are SMOKE, and baselines from another
platform_id/profile are refused.
Add workloads.json, expected digests for nine existing workloads, a dedicated
fixed-node workflow (manual, self-hosted, requires --cpu-list), and
docs/performance.md. bench.py stays as a compatibility shim. This is step 5 of
the CI/security RFC (B08).
Add tools/ci/strict-evidence.py, which records only measured host/source
facts: tested SHA (git-verified when a checkout exists), source archive
digest, kernel, bwrap/rustc/cargo, uid, cgroup2, runner kind and outcome. The
lane now writes this document on every exit path, refuses a root-only PASS,
and covers the strict @knowledge-inert and forged-marker JSON contracts in
addition to the existing 7/7 adversarial matrix and positive CLI checks.
Rework security-linux.yml to run on trusted master pushes or an explicit
source_sha, bind the checkout to the tested SHA, create a git archive for the
digest, avoid PR execution and upload the evidence artifact. This is step 4 of
the CI/security RFC (B04/B09/B10); the runner must still be an ephemeral VM.
Add an architecture test that rejects direct marker_present/verify_active
references from the CLI and keeps the Linux-only functions out of the
marqdo-sandbox crate root. Windows cross-target cargo check now compiles the
whole workspace (macOS cross-check is blocked only by ring's Darwin C/asm
build on this Linux host; its cfg(not(linux)) path is the same one Windows
exercises).
This completes the step 3 verification for the CI/security RFC B01.
Capture the strict child through launch_strict_capture and classify the
outcome before writing anything. In --json mode a control-plane failure now
emits exactly one error document with a stable code instead of forwarding
bubblewrap stderr and leaving stdout empty; a verified child's valid JSON is
forwarded unchanged with its own exit code. Empty, partial or non-JSON child
output becomes strict.child_invalid.
Route the remaining run --json error paths (mode, grants, missing entry,
trace) through the same contract, classify sandbox launch failures with
StrictFailureKind so probe/spawn/timeout/resource failures stay
distinguishable, and add deterministic routing tests plus forged-marker CLI
coverage. Document the decision in ADR-0053 and spec/security.md 1.2.
This is step 2 of the CI/security RFC (B02/B03).
The CLI referenced Linux-only marker_present/verify_active directly, so
macOS and Windows builds failed with E0425. Add StrictChildState and
inspect_strict_child() to marqdo-sandbox: Linux performs the full identity
verification, while every other platform rejects even a forged
MARQDO_STRICT_ACTIVE marker. The Linux-only functions are no longer
re-exported from the crate root, so future callers cannot bypass the
platform boundary.
This is step 1 of the CI/security RFC (B01).
Master now contains the fast-forwarded RFC-LANG-0002 series; the former
semantics-v2-refactor branch was deleted, so only master remains.
- STATUS lists the full linear series (docs, implementation, expression
grammar, status, seven follow-up fixes) and the verification at 9c1723d:
fmt/clippy clean, 507 passed / 0 failed / 2 ignored, stdlib digest verified
- counts updated to 49 compile-negative fixtures and 507 workspace tests
across README, COMPATIBILITY, CHANGELOG, CURRENT and the migration guide
- green evidence pins code commit 9c1723d and lists the follow-up fixes;
CHANGELOG gains the follow-up correctness section
- migration inventory regenerated: 1019 classified files, 0 pending
- CURRENT.md baseline is the 40-hex 9c1723d commit required by the docs gate
Verified: docs_consistency 4/4, fmt, clippy -D warnings and the full
workspace suite green; python3 tools/pack-stdlib.py PASS.
frontmatter::has_key() is what rejects marqdo_semantics/syntax directives;
add unit coverage for a present key, an absent key and invalid YAML (which
reports no key rather than panicking).
Verified: cargo test -p marqdo-markdown and the workspace suite green
(507 passed, 0 failed, 2 ignored).
spec/mapping.md §8 requires diagnostics ordered by source position, then
code, so IDE/CI diffs are stable. analyze_source appended diagnostics by
pipeline stage (units, mapping, inference, call validation), so a later
heading error could precede an earlier payload error.
Sort the final list by (span.start, code) with span-less diagnostics last,
and pin the order with a test that mixes an indexing-stage duplicate-heading
error with an earlier mapping-stage syntax error.
Verified: clippy -D warnings clean; cargo test --workspace --locked green.
RFC-LANG-0002 allows a mode declaration only as an ATX heading prefix; a
Setext heading merely inherits the nearest ancestor mode. heading_declaration
was applied to every heading, so a Setext 'title' beginning with @exec/
@knowledge silently switched modes and could create a function.
- skip declaration parsing for Setext headings (they keep their literal name
and inherited mode); add a regression test
- remove the dead normalized_heading_name helper and its imports, which only
existed for the removed link-form call resolution
Verified: clippy -D warnings clean; cargo test --workspace --locked green
(505 passed, 0 failed, 2 ignored).
A stray end after an earlier statement was reported with the span of the
previous payload (position-1) and a confusing 'ends outside its owning
function segment' message. Report it on the end payload itself with the
standard sem.unmatched_end wording, and pin the span with a unit test.
Verified: clippy -D warnings clean and cargo test --workspace --locked
green (62 targets).
signature::detect() scanned the function body independently of heading
modes, skipping only level>=3 headings. A 'params(...)' code span in an
### @knowledge section was therefore taken as the function signature
('inspect --signature' reported it as explicit), even though that span is
inert documentation and is never executed.
Unify recognition with the executable traversal:
- MappedBody now reports the params(...) payload the mapper accepted; the
mapper already enforces Exec mode, directness, first-function-wide
position and single declaration
- signature::from_payload validates that text and rewrites default-value
spans; signature::detect is removed along with its mode-blind heuristic
- analyze uses the mapper result, so Knowledge sections cannot declare a
signature and prose/Knowledge subsections may still precede the first
Exec declaration
- tests: Knowledge params ignored; Exec params after prose/Knowledge is the
signature; spec/mapping.md states the recognition rule
Verified: clippy -D warnings clean; cargo test --workspace --locked green
(62 targets).
Follow-up cleanup of text left by the concurrent implementation:
- marqdo-markdown AST, marqdo-ir statements/branch arms, frontend units,
payload and signature module docs now describe RFC-LANG-0002 roles
(payload carriers, heading modes, if/else groups, while/end, params)
- marqdo-expression public docs list parse_payload_statement and stop
describing a Markdown-link call that no longer exists
- retired markdown.invalid_construct/markdown.inert_executable constants
are documented as decode-only history for old Analysis/Trace JSON
- drop the stale 'expected red' comment in the CLI integration test
- the dev-only .mq.md review aid no longer reports semicolons inside
string literals, and skips AST baselines / expected-invalid negatives
Verified: fmt, clippy -D warnings and cargo test --workspace --locked green.
The pre-RFC link-form block-quote call was removed by RFC-LANG-0002, but the
IR still carried Stmt::Call, CallSite and CallTarget with matching branches in
the runtime, infer, call validation, loader, analysis projection and JSON
output. No mapper produces that variant any more, so it was a second call
representation that could never be reached.
- remove Stmt::Call, CallSite and CallTarget from marqdo-ir
- remove the runtime exec_call_site path and every exhaustive match arm
- remove the 'call_site' Analysis node and its JSON helpers
- state in spec/analysis-v2.md that calls are expressions; keep v1 as history
with a note that call_site is not produced by the current language
Verified: cargo clippy --workspace --all-targets -D warnings clean and
cargo test --workspace --locked green (62 targets).
The concurrent mapper tracked the first params declaration in the Mapper (so
duplicates across heading segments were caught) but checked 'first payload'
with a per-segment vector index. A params(...) in a later ### @exec segment
was therefore accepted even after an earlier segment had executed statements,
and a duplicate could even be reported as a position error.
- track whether any statement has been lowered for the function and require
params(...) to come before it, across all heading segments
- check duplicates first, then execution order, then the in-segment index
- add negative fixtures 48 (after an earlier segment) and 49 (duplicate
across segments); update the corpus count and spec wording
Verified: cargo test --workspace --locked green (62 targets).
Update status, compatibility, changelog, README, architecture overview,
ADR status, migration guide, contributor/CI/security/release notes and
the per-file inventory from 'migration pending' to the verified state:
- language SHA ffd1cc9: fmt/clippy clean, workspace 501 passed / 0 failed
/ 2 ignored, stdlib digest regenerated and pack-verified
- green evidence report added; the prior intermediate evidence is marked
superseded
- inventory regenerated: 1015 classified files, 0 pending
- explicit non-claims: Windows/macOS native lanes and the disposable-VM
strict lane were not re-run at this SHA, so the release decision remains
DO NOT RELEASE
Verified: cargo test -p marqdo-cli --test docs_consistency (4/4) and
git diff --check.
Move keyword recognition out of the Markdown mapper and into the single
expression/statement parser:
- add parse_payload_statement + PayloadStatement for if/else/while/end/
return/params/ordinary; keyword matching uses complete tokens, requires
whitespace before a control expression, rejects trailing tokens/colons
and reserves the statement words in expression positions
- keep expression spans payload-relative so the frontend can rewrite them
onto the file; malformed params(...) interiors stay owned by the
signature parser (one sem.syntax diagnostic, not a duplicate)
- simplify the frontend mapper to control-stack lowering and mapped
semantic diagnostics only
- emit sem.no_entry_function from run when no Exec H2 main exists
- add else-under-while, params-in-quote, Knowledge-inertness and
no-entry tests; update the corpus counts
- make the EBNF parameter type annotation optional (params(n) is valid)
Verified: cargo fmt --check, cargo clippy -D warnings and
cargo test --workspace --locked: 501 passed, 0 failed, 2 ignored.
One Markdown-to-program mapping, no version selector or legacy mapper.
Frontend/runtime:
- resolve @exec/@knowledge heading declarations with nearest-ancestor
inheritance; only an Exec H2 is callable; add UnitKind::Knowledge and
sem.* diagnostics (mode marker, unit name, exec-without-function,
payload nesting/empty, syntax, control stack, params position,
version directive, inferred-signature lint)
- extract Strong/Emph/inline Code as one-statement carriers, refuse
nested carriers, and honour heading segment boundaries so an if/while
cannot be closed across a heading
- lower if/else/while/end to the existing Branch/Loop/Return IR with a
compile-time control stack; keep payload-relative source spans through
expression parsing and rewrite them onto the file
- accept params(...) only as the first direct Exec H2 payload; keep the
existing free-variable inference and named-only rule for inferred
signatures; fix duplicate-mode/else/end and Knowledge-body inertness
- reject marqdo_semantics/syntax front matter directives
Data and tests:
- migrate stdlib/**/*.mq.md, regenerate stdlib.json digest
- migrate examples and the full semantics/negative/runtime corpus to
the only current syntax; expected values/runtime codes preserved
- migrate Rust-embedded .mq.md fixtures, init template, CLI help and
security-lane fixtures
Verified: cargo fmt --check, cargo clippy -D warnings and
cargo test --workspace --locked (green).
Replace the Strong/Emph/List/Quote mapping in place with the single
@exec/@knowledge heading-mode model, unified payload carriers and keyword
control flow.
- add the authoritative RFC, ADR-0052 and a historical ADR index
- rewrite spec/mapping.md and spec/expression.ebnf as the normative
mapping/statement grammar; update evaluation, references, math,
foreign, package-format, plugin-abi, security and schema notes
- rewrite README, AGENTS, COMPATIBILITY and CHANGELOG for the breaking
language change; update the architecture index/overview and active docs
- mark superseded planning docs, compute RFCs, evidence and syntax-role
ADRs as historical/superseded without rewriting their decisions
- add migration guide, per-file inventory and dev-only audit scripts
The frontend/stdlib migration is completed in the following commits; this
commit contains documentation only.