lume

Track L acceptance matrix

Schema: lume-acceptance-matrix/0.1. Source: docs/33-build-anything-plan.md.

Track L claims only. A PASS applies to the exact scoped behavior and cited command; it does not accept the umbrella issue or imply production readiness, independent verification, or coverage of another claim.

Result semantics

PASS: the cited runnable evidence supports the exact, narrowly scoped claim.
FAIL: cited runnable evidence contradicts the claim.
UNKNOWN: evidence is missing, insufficient, not independently checked where required, or not established for the current revision.
Implementation state (shipped, partial, model, or planned) is separate from evidence result.

Current declared results: 16 PASS, 0 FAIL, 8 UNKNOWN. These counts are not a maturity score.

Claim ID	Implementation	Result	Scoped claim	Evidence	Tracking
`L.ASYNC.EPOLL_READINESS`	partial	PASS	Linux epoll readiness and pipe wakeup exchange values through the native runtime.	`track-l-async-native`	#257; #318
`L.ASYNC.MULTIWORKER_EXECUTION`	partial	UNKNOWN	Multiple actual workers execute async work concurrently with work stealing.	none	#257; #318
`L.ASYNC.TIMERS_BACKPRESSURE`	model	PASS	The in-language reactor model orders timer deadlines and rejects work beyond its bounded in-flight limit.	`track-l-async-native`	#257; #318
`L.DB.BTREE_RECURSIVE_SPLIT`	partial	UNKNOWN	B+Tree insertion recursively splits full internal nodes and propagates separators through arbitrary tree heights.	none	#257; #318
`L.DB.BTREE_SHALLOW_MODEL`	model	PASS	The in-memory two-child index model inserts ordered rows, performs a shallow split, and returns range-scan results.	`track-l-db-native`	#257; #318
`L.DB.DURABLE_ACID_RECOVERY`	partial	UNKNOWN	Transactions provide durable ACID behavior with a persistent WAL and correct recovery after process crashes.	none	#257; #318
`L.DB.INMEMORY_MVCC`	model	PASS	The in-memory database model preserves snapshot reads and rolls back uncommitted changes.	`track-l-db-native`	#257; #318
`L.DB.QUERY_PLANNER`	partial	PASS	The cost-based query planner selects the pinned point-seek and range-scan plans for the tested statistics.	`track-l-db-native`	#257; #318
`L.GUI.NATIVE_WINDOWING`	planned	UNKNOWN	Applications create native OS windows and deliver real window events to Lume code.	none	#257; #318
`L.GUI.SOFTWARE_RENDERING`	partial	PASS	The software GUI model rasterizes RGBA canvases, blends pixels, solves tested layouts, dispatches hit-tested clicks, and encodes PPM/SVG frames.	`track-l-gui-native`	#257; #318
`L.MEDIA.RLE_GOP`	partial	PASS	The custom media codec round-trips the tested lossless RLE data and I/P-frame GOP residuals.	`track-l-media-native`	#257; #318
`L.MEDIA.WAV_FIR`	partial	PASS	The RIFF/WAV PCM16 codec round-trips the tested samples and the three-tap FIR filter produces the pinned output.	`track-l-media-native`	#257; #318
`L.ML.HMC`	partial	PASS	Hamiltonian Monte Carlo with leapfrog integration returns the pinned posterior summary for the tested Beta-Bernoulli model.	`track-l-ml-native`	#257; #318
`L.ML.ONNX_BINARY`	partial	UNKNOWN	ONNX export and import produce valid binary ONNX files verified by an independent parser, including adversarial malformed inputs.	none	#257; #318
`L.ML.SAFETENSORS_BINARY`	partial	UNKNOWN	SafeTensors export and import produce valid binary SafeTensors files verified by an independent parser, including adversarial malformed inputs.	none	#257; #318
`L.ML.SYMBOLIC_AUTODIFF`	partial	PASS	Compiler symbolic differentiation handles the documented polynomial and relu/sq/cube/abs/clamp/leaky_relu/quotient fragment with the tested values and gradients.	`track-l-ml-native`	#257; #318
`L.ML.WENGERT_AUTODIFF`	partial	PASS	Reverse-mode Wengert-tape differentiation produces the pinned values and gradients for the tested tanh/sigmoid/exp graph.	`track-l-ml-native`	#257; #318
`L.NET.HTTP11_LOOPBACK`	partial	PASS	The HTTP/1.1 parser/serializer and native Linux loopback client/server exchange the pinned request and response.	`track-l-net-native`	#257; #318
`L.NET.IPV4_CIDR`	partial	PASS	The native networking helpers handle the tested IPv4 and CIDR routing cases.	`track-l-net-native`	#257; #318
`L.NET.TCP_CHECKSUM_COMPLETE`	partial	UNKNOWN	TCP checksum generation and verification cover the complete TCP pseudo-header, header, payload, odd-byte padding, and carry rules with independent vectors.	none	#257; #318
`L.NET.TCP_FSM_MODEL`	model	PASS	The in-language TCP state model reaches ESTABLISHED through the tested handshake transitions.	`track-l-net-native`	#257; #318
`L.NET.TLS13_INTEROP`	partial	UNKNOWN	TLS 1.3 clients and servers interoperate with an independent implementation using standard cryptography, certificate handling, and adversarial protocol tests.	none	#257; #318
`L.NET.UDP_CHECKSUM`	partial	PASS	The tested UDP packet helper computes an RFC 1071-style checksum and rejects modified payloads and lengths.	`track-l-net-native`	#257; #318
`L.NET.UDP_LOOPBACK`	partial	PASS	Native Linux UDP loopback sends and receives the pinned datagram.	`track-l-net-native`	#257; #318

Sources, evidence, and boundaries

`L.ASYNC.EPOLL_READINESS`

Source: compiler/stage1/async_rt.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-async-native:

make -C bootstrap/lume0 test

Expected: The native conformance output contains the pinned epoll pipe values and the async_rt unit tests pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: This is a Linux readiness increment, not a claim of multi-worker execution.

`L.ASYNC.MULTIWORKER_EXECUTION`

Source: compiler/stage1/async_rt.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: The existing scheduler processes list state sequentially. No runnable evidence demonstrates actual concurrent workers.

`L.ASYNC.TIMERS_BACKPRESSURE`

Source: compiler/stage1/async_rt.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-async-native:

make -C bootstrap/lume0 test

Expected: The pinned timer result and rejected submission count match, and async_rt unit tests pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS covers the tested reactor model only.

`L.DB.BTREE_RECURSIVE_SPLIT`

Source: compiler/stage1/db.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: The current two-child node split does not establish recursive internal-node splitting. Existing shallow tests are insufficient.

`L.DB.BTREE_SHALLOW_MODEL`

Source: compiler/stage1/db.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-db-native:

make -C bootstrap/lume0 test

Expected: The db unit tests and pinned height, point-read, and range-sum observations pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/db.src, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: This scoped model PASS does not establish recursive splits, durable pages, or a complete B+Tree.

`L.DB.DURABLE_ACID_RECOVERY`

Source: compiler/stage1/db.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: In-memory transaction state and WAL string serialization are not durable storage or crash-recovery evidence.

`L.DB.INMEMORY_MVCC`

Source: compiler/stage1/db.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-db-native:

make -C bootstrap/lume0 test

Expected: The pinned snapshot, committed value, and rollback observations match, and db unit tests pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/db.src, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS is limited to in-memory snapshot and rollback behavior; it is not durable ACID acceptance.

`L.DB.QUERY_PLANNER`

Source: compiler/stage1/db.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-db-native:

make -C bootstrap/lume0 test

Expected: The native output selects IndexPointSeek and IndexRangeScan for the tested queries.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS covers only the tested plan choices.

`L.GUI.NATIVE_WINDOWING`

Source: compiler/stage1/gui.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: Software framebuffers and SVG/PPM encoders do not create native windows. No runnable native-window evidence exists.

`L.GUI.SOFTWARE_RENDERING`

Source: compiler/stage1/gui.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-gui-native:

make -C bootstrap/lume0 test

Expected: The gui unit tests and pinned pixel, blend, click, and serialization observations pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/gui.src, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS does not include native window creation or OS event integration.

`L.MEDIA.RLE_GOP`

Source: compiler/stage1/media.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-media-native:

make -C bootstrap/lume0 test

Expected: The media unit tests and pinned GOP run count and decoded frame pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/media.src, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS is limited to Lume’s custom codec and does not claim standards-based video interoperability.

`L.MEDIA.WAV_FIR`

Source: compiler/stage1/media.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-media-native:

make -C bootstrap/lume0 test

Expected: The media unit tests and pinned WAV decode, sample-rate, and FIR observations pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/media.src, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS covers the tested PCM16 and FIR behavior only.

`L.ML.HMC`

Source: compiler/stage1/ai_runtime.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-ml-native:

make -C bootstrap/lume0 test

Expected: The native HMC observation falls within the pinned tested range.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS covers one tested inference path, not general probabilistic-programming correctness.

`L.ML.ONNX_BINARY`

Source: compiler/stage1/nn.src, compiler/stage1/track_l_domains_ok.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: Existing descriptors and self-verification do not provide independent binary ONNX parser acceptance.

`L.ML.SAFETENSORS_BINARY`

Source: compiler/stage1/nn.src, compiler/stage1/track_l_domains_ok.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: Existing descriptors and substring checks do not provide independent binary SafeTensors parser acceptance.

`L.ML.SYMBOLIC_AUTODIFF`

Source: compiler/stage1/lume/06b_phase3.src, compiler/stage1/lcaas.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-ml-native:

make -C bootstrap/lume0 test

Expected: The pinned values and gradients match, and the tested contracts are reported proven.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS is restricted to the documented fragment and test observations.

`L.ML.WENGERT_AUTODIFF`

Source: compiler/stage1/nn.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-ml-native:

make -C bootstrap/lume0 test

Expected: The native output matches the pinned tape values and both gradients.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS covers the tested graph, not arbitrary tensor programs.

`L.NET.HTTP11_LOOPBACK`

Source: compiler/stage1/net.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-net-native:

make -C bootstrap/lume0 test

Expected: The net unit tests and pinned native HTTP status, method, path, and response body pass.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/net.src, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: This loopback increment is not evidence of a production HTTP service or TLS interoperability.

`L.NET.IPV4_CIDR`

Source: compiler/stage1/net.src.

Evidence track-l-net-native:

make -C bootstrap/lume0 test

Expected: The net.src unit tests, including IPv4/CIDR cases, pass in the full suite.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/net.src.

Boundary: PASS covers existing unit cases only.

`L.NET.TCP_CHECKSUM_COMPLETE`

Source: compiler/stage1/net.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: The current segment-word helper is not complete TCP checksum evidence and lacks independent vectors for the full requirements.

`L.NET.TCP_FSM_MODEL`

Source: compiler/stage1/net.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-net-native:

make -C bootstrap/lume0 test

Expected: The pinned handshake observation reaches ESTABLISHED.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS is for a state model only, not complete TCP behavior.

`L.NET.TLS13_INTEROP`

Source: compiler/stage1/net.src, compiler/stage1/track_l_domains_ok.src.

Evidence: UNKNOWN — no sufficient runnable acceptance evidence is registered.

Boundary: Custom key-schedule and record routines tested only against Lume itself are not interoperable TLS 1.3 evidence.

`L.NET.UDP_CHECKSUM`

Source: compiler/stage1/net.src.

Evidence track-l-net-native:

make -C bootstrap/lume0 test

Expected: The net.src RFC 1071 checksum and UDP tamper-detection unit test passes.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/net.src.

Boundary: PASS does not promote the incomplete TCP checksum or establish full UDP interoperability.

`L.NET.UDP_LOOPBACK`

Source: compiler/stage1/net.src, compiler/stage1/track_l_domains_ok.src.

Evidence track-l-net-native:

make -C bootstrap/lume0 test

Expected: The native output contains the pinned UDP loopback payload.

Artifacts: bootstrap/lume0/tests/run.sh, compiler/stage1/track_l_domains_ok.src, tests/conformance/stage1/track_l_domains_ok_asm.expected.

Boundary: PASS is limited to the tested Linux loopback exchange.

Consumer commands

./lume build acceptance/matrix_assertions.src -o build/bin/acceptance_matrix
printf '@mode local\n@root %s\n' "$PWD" | build/bin/acceptance_matrix
printf '@mode online\n@root %s\n' "$PWD" | build/bin/acceptance_matrix

The local command validates artifacts and negative fixtures and checks this generated view. The online command additionally verifies issue/PR existence and expected state through the GitHub API. Runnable evidence commands are executed by the existing project test jobs; this validator does not treat a declared command as an executed result.

This site is open source. Improve this page.