Files
reverse-engineering-anything/src/application/javascript/JavaScriptArtifactGraphBuilder.ts
T
ShuaixinHuang e9fc8fb30c fix(javascript): preserve runtime path bytes and Electron ownership (#4941)
* fix(javascript): preserve literal percent runtime paths

* fix(javascript): ignore resolve prefixes before the final anchor

* fix(electron): retain application roots for loadFile declarations

* fix(electron): resolve loadFile from enclosing application package

* fix(electron): retain outer application root across package scopes

* fix(electron): derive application ownership from inventoried packages

* fix(electron): retain resolved package entrypoint ownership

* fix(electron): preserve competing application ownership

* fix(electron): retain invalid and omitted runtime ownership

* fix(electron): preserve omitted module ownership and runtime paths

* fix(electron): preserve proven roots before confining page paths

* fix(electron): retain possible owners behind missing imports

* fix(electron): distinguish absent imports and owned page candidates

* fix(electron): keep unavailable JSON out of owner gaps

* fix(electron): retain exact loader formats and default entry status

* fix(electron): retain unavailable entrypoint ownership candidates

* test(ghidra): preserve Windows Node probe prerequisites

* test(ghidra): assert cancellation during profile shutdown

* fix(electron): separate examined imports from frontend ownership

* fix(electron): retain unavailable directory entrypoint owners

* test(jeb): retain Linux ownership prerequisite in isolation

* fix(javascript): retain unchecked loader ownership and literal main paths

* fix(javascript): preserve runtime paths and call-site module aliases

* fix(javascript): bound module alias traversal and preserve projections

* fix(javascript): retain host module and nested artifact path boundaries

* fix(javascript): preserve ESM and archive module loader provenance

* test(javascript): distinguish unmapped containers from missing exports

* fix: retain default module aliases and Electron builtin imports

* fix: preserve host file URL and inward ASAR ownership

* fix: preserve CommonJS file URL classification in CI fixtures

* fix: retain bound module loaders and Windows entrypoint uncertainty

* fix: retain ASAR uncertainty and invalidate changed Electron app roots

* fix(javascript): retain ownership uncertainty across loader boundaries

* fix(javascript): preserve loader sources and bound alias queries

* fix(javascript): retain destructured loader sources

* fix(javascript): preserve loader uncertainty at resolution boundaries

* fix(javascript): preserve callable and package launch ownership

* fix(javascript): preserve captured values across member reads

* test(process): assert tagged filesystem cancellation failures

* fix(javascript): preserve inventory bases and invocation read timing

* fix(javascript): preserve ownership across indirect and unread loads

* test(javascript): order fixture directory creation on Windows

* test(javascript): share ordered filesystem fixture writer

* fix(javascript): unify loader provenance and package validation

* fix(javascript): preserve captured Node require loader values

* fix(javascript): preserve loader invocation authority

* fix(javascript): verify generic loader invocation authority

* fix(javascript): preserve conditional loader method alternatives

* fix(javascript): require proven factories for source association

* fix(javascript): honor replaced Electron app path setters

* fix(javascript): prove path and assignment source ownership

* fix(javascript): retain native delegate eligibility through shared writes

* fix(process): settle stale baseline errors after confirmed absence

* fix(javascript): preserve assignment and path boundary evidence

* fix(javascript): prove source-directory resolve anchors

* fix: preserve application path provenance and portable schemas

* fix(javascript): retain CommonJS exported setter effects

* fix: isolate Linux identity observations and retain native module exports

* fix(javascript): preserve published app and helper provenance

* fix(javascript): normalize typed stores and native path fixtures
2026-10-20 17:25:19 +08:00

396 lines
15 KiB
TypeScript

import { ELECTRON_IDENTITY_LIMITATION } from "../../domain/javascript/javascriptElectronMemberWrites.js";
import type { ArtifactInventorySnapshot } from "../../domain/artifactInventorySnapshot.js";
import {
createJavaScriptApplicationGraph,
sealTransferredJavaScriptApplicationGraphSteps,
type JavaScriptApplicationGraph,
} from "../../domain/javascript/javascriptApplicationGraph.js";
import type { JavaScriptModuleArtifactAnalysis } from "./JavaScriptArtifactAnalysisTypes.js";
import type { JavaScriptArtifactFileSet } from "../../domain/javascript/javascriptArtifactFiles.js";
import { JavaScriptArtifactGraphAccumulator } from "./JavaScriptArtifactGraphAccumulator.js";
import {
HTML_INLINE_SCRIPT_LOCATION_COVERAGE_LIMIT,
omittedHtmlInlineScriptLocations,
selfReferenceOmissions,
dependencyManifest,
applicationContainerIdentity,
type JavaScriptArtifactGraphContext,
} from "./JavaScriptArtifactGraphContext.js";
import {
addJavaScriptHtmlRoles,
addJavaScriptSourceMapOriginals,
} from "./JavaScriptArtifactGraphDocuments.js";
import { addJavaScriptStaticFindingsSteps } from "./JavaScriptArtifactGraphFindings.js";
import {
addJavaScriptModuleRelationshipsSteps,
type JavaScriptModuleRelationshipOmissions,
addJavaScriptSourceModules,
} from "./JavaScriptModuleRelationships.js";
import {
completeApplicationCoverage,
partialApplicationCoverage,
truncatedApplicationCoverage,
} from "../../domain/javascript/javascriptApplicationEvidenceSchemas.js";
import type { JavaScriptSemanticResourceLimit } from "../../domain/javascript/javascriptSemanticValueTypes.js";
import {
semanticCoverageResourceLimits,
semanticResourceLimitCoverage,
} from "../../domain/javascript/javascriptSemanticCoverage.js";
import { semanticResourceLimitReason } from "../../domain/javascript/javascriptSemanticResourceLimits.js";
import {
addJavaScriptArtifactContainers,
addJavaScriptArtifactFiles,
addJavaScriptPackageNodes,
createJavaScriptArtifactRootNode,
} from "./JavaScriptArtifactGraphStructure.js";
import { addJavaScriptBundlerNodes } from "./JavaScriptArtifactGraphBundlers.js";
import { addElectronBoundaries } from "./ElectronBoundaryGraph.js";
import {
classifyElectronIpcPairings,
collectElectronIpcRecords,
} from "./ElectronBoundaryAnalysis.js";
/** Project artifact and AST facts into JavaScript Application Graph. */
export const buildJavaScriptArtifactGraph = (
snapshot: ArtifactInventorySnapshot,
fileSet: JavaScriptArtifactFileSet,
analysis: JavaScriptModuleArtifactAnalysis,
): JavaScriptApplicationGraph =>
createJavaScriptApplicationGraph(
completeSynchronously(
buildJavaScriptArtifactGraphInputSteps(snapshot, fileSet, analysis),
),
);
/**
* Build a validated application graph with cooperative immutable ownership.
* Constructing the graph input also runs in steps: projecting every module
* relationship and finding at once held the event loop for seconds.
*/
export function* buildImmutableJavaScriptArtifactGraphSteps(
snapshot: ArtifactInventorySnapshot,
fileSet: JavaScriptArtifactFileSet,
analysis: JavaScriptModuleArtifactAnalysis,
): Generator<void, JavaScriptApplicationGraph> {
const input = yield* buildJavaScriptArtifactGraphInputSteps(
snapshot,
fileSet,
analysis,
);
// The input was built here and is not retained, so it can be transferred.
return yield* sealTransferredJavaScriptApplicationGraphSteps(input);
}
const completeSynchronously = <Value>(steps: Iterator<void, Value>): Value => {
for (;;) {
const step = steps.next();
if (step.done === true) return step.value;
}
};
function* buildJavaScriptArtifactGraphInputSteps(
snapshot: ArtifactInventorySnapshot,
fileSet: JavaScriptArtifactFileSet,
analysis: JavaScriptModuleArtifactAnalysis,
): Generator<void, unknown> {
const accumulator = new JavaScriptArtifactGraphAccumulator();
const root = createJavaScriptArtifactRootNode(accumulator, snapshot);
const parsedPackages = new Set(
analysis.packages
.filter((metadata) => metadata.status === "included")
.map(({ path }) => path),
);
const applicationOwnershipUnknownContainers = new Set<string>();
const applicationModuleGapSources = new Map<string, Set<string>>();
const applicationModuleReferencesRetained = new Set<string>();
const recordModuleGap = (file: {
path: string;
container_sha256: string;
}) => {
const container = applicationContainerIdentity(fileSet, file);
const sources =
applicationModuleGapSources.get(container) ?? new Set<string>();
sources.add(file.path);
applicationModuleGapSources.set(container, sources);
};
for (const files of [fileSet.files, fileSet.unavailable_relevant_files]) {
for (const file of files) {
if (
file.kind === "package-json" &&
!dependencyManifest(file.path) &&
!parsedPackages.has(file.path)
)
applicationOwnershipUnknownContainers.add(
applicationContainerIdentity(fileSet, file),
);
if (
file.kind === "javascript" &&
(!("text" in file) || !file.text.included)
)
recordModuleGap(file);
}
}
// Omitted module relationships can hide another launch owner. Do not expand
// retained facts beyond the projection guard merely to recover ownership.
for (const analyzed of analysis.files) {
if (
analyzed.application_projection_failure !== undefined ||
analyzed.analysis_failure !== undefined ||
analyzed.javascript?.parse_status === "failed" ||
(analyzed.javascript?.parse_error_count ?? 0) > 0
)
recordModuleGap(analyzed.file);
else if (analyzed.javascript !== null)
applicationModuleReferencesRetained.add(analyzed.file.path);
}
const unavailableFilesByPath = new Map(
fileSet.unavailable_relevant_files.map((file) => [file.path, file]),
);
const context: JavaScriptArtifactGraphContext = {
accumulator,
snapshot,
fileSet,
analysis,
directoryPaths: new Set(
snapshot.occurrences
.filter(({ entry_kind }) => entry_kind === "directory")
.map(({ logical_path }) => logical_path),
),
root,
filesByPath: new Map(fileSet.files.map((file) => [file.path, file])),
unanalyzedFiles: {
get: (path) =>
unavailableFilesByPath.get(path) ?? fileSet.unanalyzed_files?.get(path),
hasDirectory: (path, container) =>
fileSet.unanalyzed_files?.hasDirectory(path, container) ?? false,
},
fileNodes: new Map(),
assetNodes: new Map(),
sourceModuleNodes: new Map(),
chunkNodes: new Map(),
moduleNodes: new Map(),
containerNodes: new Map([[snapshot.manifest.root_sha256, root]]),
applicationRootsBySource: new Map(),
applicationOwnershipUnknownContainers,
applicationModuleGapSources,
applicationModuleReferencesRetained,
applicationPossibleRootsByContainer: new Map(),
};
addJavaScriptArtifactContainers(context);
yield;
addJavaScriptArtifactFiles(context);
yield;
// Raw observations remain on the file nodes. Only admitted expansions feed
// the relationship builders, including their Electron and bundler callers.
const projectionContext = {
...context,
analysis: {
...analysis,
files: analysis.files.map((file) =>
file.application_projection_failure === undefined
? file
: { ...file, javascript: null, semantic: null },
),
},
};
const packageRoots = addJavaScriptPackageNodes(projectionContext);
yield;
addJavaScriptSourceModules(projectionContext);
yield;
const bundlerLimitations = addJavaScriptBundlerNodes(projectionContext);
const relationshipOmissions =
yield* addJavaScriptModuleRelationshipsSteps(projectionContext);
const findingLimitations =
yield* addJavaScriptStaticFindingsSteps(projectionContext);
addElectronBoundaries(projectionContext);
yield;
addJavaScriptHtmlRoles(projectionContext);
yield;
addJavaScriptSourceMapOriginals(projectionContext);
yield;
const coverage = graphCoverage(context);
return {
schema: "JavaScriptApplicationGraph",
root_node_ids:
packageRoots.length === 0
? [root.node_id]
: packageRoots.map(({ node_id: id }) => id),
nodes: accumulator.nodes(),
edges: accumulator.edges(),
coverage,
limitations: [
...bundlerLimitations,
...findingLimitations,
...graphLimitations(context, coverage.status, relationshipOmissions),
],
};
}
const graphCoverage = (context: JavaScriptArtifactGraphContext) => {
const resourceLimits = semanticResourceLimits(context);
const sourceMapPolicyGap = context.analysis.source_maps.some(
({ status }) => status === "invalid",
);
const malformedStructuredData =
context.analysis.packages.some(({ status }) => status !== "included") ||
context.analysis.json_modules.some(({ status }) => status !== "included") ||
context.analysis.source_maps.some(({ status }) => status === "invalid");
const partialJavaScript = context.analysis.files.some(
({ javascript }) =>
javascript !== null && javascript.parse_status === "partial",
);
const unknownGap =
(context.fileSet.source_selection?.unselected_files ?? 0) > 0 ||
context.fileSet.unavailable_relevant_files.length > 0 ||
context.analysis.files.some(
({ analysis_failure, application_projection_failure }) =>
analysis_failure !== undefined ||
application_projection_failure !== undefined,
) ||
context.analysis.parse_failures > 0 ||
context.fileSet.invalid_utf8_files > 0 ||
context.analysis.html_inline_scripts.length > 0 ||
sourceMapPolicyGap ||
malformedStructuredData ||
partialJavaScript;
const executionLimits = context.analysis.files.flatMap(
({ analysis_failure, application_projection_failure }) =>
[analysis_failure, application_projection_failure].flatMap((failure) =>
failure === undefined
? []
: javaScriptAnalysisCoverageLimits(failure.limits),
),
);
// Capped inline script locations truncate the result. The total omission
// stays unknown because the unanalyzed bodies hide an unknown fact count.
const inlineScriptsTruncated =
omittedHtmlInlineScriptLocations(context.analysis.html_inline_scripts) > 0;
if (
resourceLimits.length > 0 ||
executionLimits.length > 0 ||
inlineScriptsTruncated
) {
const limits = [
...new Map(
[
...semanticResourceLimitCoverage(resourceLimits),
...executionLimits,
...(inlineScriptsTruncated
? [HTML_INLINE_SCRIPT_LOCATION_COVERAGE_LIMIT]
: []),
].map((limit) => [JSON.stringify(limit), limit]),
).values(),
];
return inlineScriptsTruncated
? truncatedApplicationCoverage(limits, null)
: partialApplicationCoverage(limits, null);
}
if (unknownGap) return partialApplicationCoverage([], null);
if (context.snapshot.integrity_contradictions.length > 0) {
const nestedArchiveWasOpaque =
context.snapshot.integrity_contradictions.some(({ logical_path }) =>
logical_path.toLowerCase().endsWith(".asar"),
);
return partialApplicationCoverage([], nestedArchiveWasOpaque ? null : 0);
}
return completeApplicationCoverage();
};
const semanticResourceLimits = (
context: JavaScriptArtifactGraphContext,
): JavaScriptSemanticResourceLimit[] =>
[
...new Set(
context.analysis.files.flatMap(({ semantic }) =>
semantic === null
? []
: semanticCoverageResourceLimits(semantic.ir.coverage),
),
),
].sort();
const graphLimitations = (
context: JavaScriptArtifactGraphContext,
coverage: "complete" | "partial" | "unknown" | "unavailable",
relationshipOmissions: JavaScriptModuleRelationshipOmissions,
): string[] => {
const ipc = collectElectronIpcRecords(context.analysis);
const pairings = classifyElectronIpcPairings(ipc);
const electronFindings = context.analysis.files.reduce(
(count, { javascript }) =>
count +
(javascript === null
? 0
: javascript.electron.browser_windows.length +
javascript.electron.context_bridge_apis.length +
javascript.electron.ipc.length +
javascript.electron.sender_validations.length +
javascript.electron.utility_processes.length +
javascript.electron.native_addon_bindings.length),
0,
);
return [
...context.analysis.limitations,
...(context.fileSet.source_selection === undefined
? []
: [
`Caller-selected source analysis leaves ${String(context.fileSet.source_selection.unselected_files)} inventoried files outside analysis scope. Inventory identity covers the complete artifact; relationships through unselected sources remain unknown.`,
]),
...context.snapshot.integrity_contradictions.map(
({ logical_path: path }) =>
`Artifact integrity metadata contradicts observed bytes at ${path}; the observed bytes are untrusted.`,
),
...selfReferenceOmissions(
relationshipOmissions.selfImports,
["import specifier", "import specifiers"],
"the importing module",
),
"CommonJS and ESM binding relationships were recovered from inert syntax and resolved only within the inventoried artifact container.",
"Webpack/Rspack factories were recovered from AST literals; REA did not invoke push handlers or bundle bootstrap code.",
"Static imports, entrypoints, workers, endpoints, and storage relationships do not prove runtime execution.",
...(context.analysis.files.some(({ javascript }) =>
javascript?.limitations.includes(ELECTRON_IDENTITY_LIMITATION),
)
? [ELECTRON_IDENTITY_LIMITATION]
: []),
...(electronFindings === 0
? []
: [
"Electron relationships are derived from inert syntax; runtime registration, reachability, defaults, and enforcement remain unproven.",
]),
...(ipc.some(({ finding }) => finding.channel === null)
? [
"Dynamic IPC channel expressions remain unresolved and are never paired by textual similarity.",
]
: []),
...(pairings.some(({ status }) => status === "ambiguous")
? [
"Some literal IPC channels have multiple compatible main handlers; ambiguous calls are not paired to any handler.",
]
: []),
...(context.analysis.files.some(
({ javascript }) =>
(javascript?.electron.sender_validations.length ?? 0) > 0,
)
? [
"Sender, frame, URL, and origin checks are observations only; REA does not claim that they enforce a complete authorization policy.",
]
: []),
...semanticResourceLimits(context).map(semanticResourceLimitReason),
...(context.analysis.files.some(
({ javascript }) =>
(javascript?.electron.native_addon_bindings.length ?? 0) > 0,
)
? [
"Native member names are requested by JavaScript syntax and are not verified binary exports in this workflow.",
]
: []),
...(coverage === "complete"
? []
: [
"Graph coverage is incomplete; unavailable facts remain explicit and must not be read as absence.",
]),
];
};
import { javaScriptAnalysisCoverageLimits } from "../../domain/javascript/javascriptAnalysisResourceControls.js";