Web Forms · reproducible local demo

Replay the public corpus, then inspect the right artifact.

The one-command local workflow builds checked-in synthetic Web Forms projects, scans attached and separate provider layouts, retains bounded reports, and refuses unsupported or incomplete acceptance states. It does not need customer code or a database.

Public claim level: demo. The validation receipt below is bound to one exact public synthetic tree and durable Windows result. Workflow orientation without a matching projected artifact remains concept. Raw indexes, facts, logs, test receipts, binaries, source, and machine-local paths remain hidden.

Choose the replay

Local static replay and Windows publication are separate gates.

Default local replayUse PowerShell 7, the documented .NET toolchain, and a fresh output root. pwsh -File scripts/wlocal.ps1 -OutputRoot <fresh-output-root> exercises checked-in public fixtures only.

Authentic publication-RequireWindowsPublish additionally requires mapped and mapless ASP.NET publication on Windows. A non-Windows refusal or skipped Windows theory is not a pass.

PrerequisitesFollow the validation guide, the native workflow, and the public corpus map.

Execution boundaryThe workflow builds and analyzes fixture artifacts. It does not launch a Web Forms site, execute fixture database methods, run SQL, or observe a user request.

Exact-tree recipe

Start clean and preserve failures.

Check out the exact tested head shown in the receipt, install the documented prerequisites, choose a new output root, and keep the result folder when a gate refuses admission. Later trees need their own result; this count is not current-head proof.

git clone https://github.com/joefeser/tracemap.git
cd tracemap
git checkout afe6152ccb58ad11b61ef4782ba82afdfe7651b7
pwsh -File scripts/wlocal.ps1 -OutputRoot <fresh-output-root>

Artifact map

Each retained class answers one bounded question.

The map describes roles only. Public review uses the allowlisted projection; retained raw artifacts stay local.

Retained artifact classes and their public treatment.
Artifact class	Question answered	Public treatment	Boundary
Validation receipt	Which bounded generator, inputs, test result, platform, and acceptance mode were admitted?	Publish only categorical counts, rule/tier, durable result, and separately labeled source-receipt hashes.	The local-only receipt and its bounded input roster are not published.
Test receipt	Did the named synthetic suite pass, fail, or skip?	Summarize exact pass/fail/skip counts from the admitted receipt.	Raw TRX remains hidden; a skipped Windows case is not a pass.
Scan/report family	Which source or compiled scope was indexed?	Use the allowlisted concept projection.	No raw `facts.ndjson`, SQLite, logs, or report bodies.
Combined artifact	Which independent website and provider scopes were available?	Describe labels and source/compiled roles.	The combined index remains local and does not prove runtime composition.
Path view	Which ordered static route or explicit gap was retained?	Show bounded labels, rules, tiers, coverage, unresolved state, and gaps through the projection.	No source snippets, command bodies, parameter values, or database identity.
Cap comparison	What disappeared at the configured one-path result bound?	Show the admitted count and `TruncatedByLimit` gap under `combined.paths.truncation-gap.v1`, Tier4Unknown.	One retained path is not complete application coverage.
Unresolved-value view	Which static value stayed unknown?	Show a categorical unresolved operand and owner question.	No raw SQL, procedure name, scheduled command, or connection material.

Five layouts

Input placement changes the evidence available, not the claim boundary.

Public synthetic operator layouts.
Layout	Admitted inputs	Expected review result	Claim
`attached`	Website source plus website and provider DLLs.	Three bounded terminal candidates, including the separately projected Fill route.	`demo` layout pass; projected path details remain `concept`.
`separate`	Website scan and provider scan combined after independent indexing.	The same admitted getter/constructor/profile/provider route shape.	No runtime service or provider selection claim.
`separate-dll-only`	Provider scan excludes VB source and retains compiled provider evidence.	Three terminal routes remain available through compiled candidates.	Compiled-only candidates do not prove source freshness or execution.
`reversed`	Provider-first combine order.	The admitted static route and command-state assertions remain the same.	Order independence is bounded to this synthetic corpus. `validation.deep-projectless-corpus.v1` · Tier2Structural · `bounded-exact-tree-synthetic-layout` · exact-tree receipt.
`missing`	Website source and website DLL without the provider scope.	No invented database route; `SelectorNoMatch` and other Tier4Unknown gaps stay visible.	Missing evidence is not proof of runtime absence.

Four query views

Bounds and filters stay attached to the result.

allFor the four positive layouts, retains all three admitted static terminal candidates plus explicit gaps for the selected synthetic handler. The missing layout retains gaps and zero paths.

fillFilters to the independent Fill terminal. A zero-unresolved subset does not resolve command values excluded by the filter.

cappedRetains one path and a TruncatedByLimit Tier4Unknown gap. The cap is a coverage boundary, not a severity or completeness judgment.

repeatChecks admitted paths, gaps, and index bytes for deterministic equality. Repeat success does not prove that a prior intermittent failure was fixed. validation.deep-projectless-corpus.v1 · Tier2Structural · bounded-exact-tree-synthetic-layout · exact-tree receipt.

Selected public evidence

Reuse the projection; do not copy the retained run.

Claim level: concept. These examples come from the #806 allowlisted projection, whose illustrative hop IDs, tiers, and spans are not independent extractor-output proof.

Attached and separate

The projected Dynamic email lookup and Literal audit call preserve source, constructor, provider, and API-terminal categories while unresolved operands and branch feasibility remain explicit gaps.

combined.paths.compiled-il-bridge.v1 · Tier2Structural/Tier3SyntaxOrTextual candidates · bounded static evidence

Compiled-only provider

The projected Independent Fill terminal keeps the provider DLL boundary and database API terminal visible without claiming that a runtime implementation was selected or rows were returned.

dotnet.compiled.member.v1 · Tier2Structural metadata · review candidate

Missing and capped

Missing-provider refusal preserves query gaps instead of inventing a route. The capped view preserves TruncatedByLimit rather than presenting one path as complete coverage.

combined.paths.query-gap.v1 / combined.paths.truncation-gap.v1 · Tier4Unknown

Open the public-safe projection or read its evidence and limitation guide.

Exact-tree validation receipt

A durable Windows result, not a floating current-head claim.

Verified bindings for the admitted public synthetic result.
Field	Verified value	Meaning
Tested PR head	`afe6152ccb58ad11b61ef4782ba82afdfe7651b7`	Exact commit exercised by the durable result.
Tree-equivalent main merge	`d76358f663ce40532fe0954ca9888612d51a4121`	Shares tree `1163de1ea0bc4f57cd762ebc3a85ee014bc6fff2` with the tested head.
Durable result	GitHub Actions run 37125942534	Windows Server 2025; required mapped and mapless ASP.NET publication.
Counts	148 passed · 0 failed · 0 skipped	The test count includes 2 Windows publication cases, 2 profile dynamic-lookup cases, and 5 operator layouts. Four additional diagnostic-layout admission checks are recorded separately in the receipt.
Receipt class	`validation.deep-projectless-corpus.v1` · `Tier2Structural`	Synthetic static validation with explicit limitations.
Source-receipt generator SHA-256	`529f399c5492f89d12785beb091eda70b7ea6191be60c8fd53a81e818027ec39`	Digest of the exact local validation wrapper used by that result.
Source-receipt bounded-input SHA-256	`90120a5650f5da8472f7ffc734b05a74e9b4ef2075d123fafc35f07addc96a12`	Covers the admitted source roster, execution/fixture assemblies, and test receipt. It is not a public-projection input digest.
Public-projection generator SHA-256	`a9aae27d806b21bb1d8c48f861d0b82533e0862f1c8b12e1683ad58027f4946c`	Digest for the separate #806 public generator.
Public-projection bounded-input SHA-256	`23000bf21810695f70f3b5e9f96f460037c1611534b8a5ab7d63a8d63effcd54`	Covers only the allowlisted privacy-projected input; it is independent of the source receipt.

This exact-tree result remains valid evidence for its named tree. It is historical with respect to later main revisions and must not be relabeled as validation of this page's implementation head.

Failure remains evidence

Refusal is not converted into partial success.

Fail-closed outcomes and owner next steps.
Condition	Outcome	Owner action
Existing output root	`DEEP_CORPUS_OUTPUT_NOT_FRESH`; tests do not start.	Choose a new empty destination.
Windows publication required on another platform	`DEEP_CORPUS_WINDOWS_REQUIRED`; no output is written.	Move the exact tree to a suitable Windows runner.
Test process fails	`DEEP_CORPUS_TEST_FAILED;outputs-preserved`; no admitted validation receipt.	Inspect retained failure evidence; do not call the corpus validated.
Required layout, wizard case, or Windows publication case missing	Validation refuses admission.	Restore the exact prerequisite or record the gap.
Generator or bounded input changes	`DEEP_CORPUS_INPUT_CHANGED;outputs-preserved-not-admitted`.	Run again from a fresh root and bind the new exact tree.
Cap, repeat, comparison, or unresolved-ledger mismatch	Corpus fails rather than reporting a partial pass.	Investigate the changed assertion and retain the failed output.

Non-claims and remaining gaps

The demo stops at deterministic static evidence.

No runtime page, fixture method, database command, branch, provider dispatch, parameter value, or returned-row observation.
No customer compatibility, private application coverage, migration parity, complete cross-service tracing, physical drive-I/O measurement, release approval, or safety-to-run claim.
Candidate bridges, unresolved values, Tier4Unknown gaps, result caps, and partial/reduced coverage remain visible; missing evidence is not runtime absence.
PR #803 diagnostics and scheduling repairs are not ancestors of this page's exact main boundary and are not described as shipped.
No raw SQLite, facts.ndjson, analyzer log, TRX, source snippet, raw SQL, command body, configuration, credential, connection material, private identity, local path, customer artifact, or unpublished validation output is public.
No LLM call, embedding, vector database, or prompt-based classification is part of this evidence.

Continue from reproduction to review.

Use the guided setup for operator choices, the source + compiled proof for the projected chain, review handoff for owner questions, validation for repository-wide gates, outputs for retention boundaries, and limitations before drawing a conclusion. The later review-workbench story will add a dedicated walkthrough.