Preserve canonical distinct-source and unique-native-observation refinements after resolving inline or retained JavaScript workflow inputs. The shared application entry points now return structured invalid_request errors before graph parsing and computation, avoiding another complete graph capture.
Six focused tests, actual compiled CLI/SDK stdio workflows, JSON Schema dialect checks, static/document checks, and a real Notion ASAR analysis/recovery/trace/close workflow passed. All 20 final-head checks, including real JavaScript recovery, passed. Native-provider acceptance is not claimed for this provider-neutral validation change.
Fixes#1364
Refs #1065
Stream large CLI JSON files through a strict UTF-8 decoder and standard JSON parser using the same admitted regular-file handle. Preserve native parsing for small inputs, typed malformed-input/resource/access-denied errors, cancellation, and complete Evidence.
Verified with focused public CLI regressions, a 768 MB whitespace-padded valid request, and the Windows native decoder boundary. Individual strings and the assembled object graph remain subject to native runtime limits. All 27 final-head checks passed.
Refs #1331
Reject positively identified foreign-only Ghidra decompiler binaries and non-regular native prerequisites before provider startup, while retaining unknown classifications for incomplete or unrecognized headers.
Verified with 54 focused tests and real Linux Ghidra CLI/MCP admission, decompilation, cleanup, wrong-host PE rejection, and non-blocking FIFO rejection. All 19 final-head checks passed.
Fixes#1192
Co-authored-by: Dery Ferdika <114888155+DeryFerd@users.noreply.github.com>
BREAKING CHANGE: Filesystem digest omission entries now require system_code; budget and changed-file omissions use null, and unavailable-file omissions preserve the operating-system error code.
Reject omitted segment bounds, procedure classification and body facts, unresolved-call observations, dossier facets, and Ghidra parameter flow metadata. Providers emit explicit nulls for unsupported facets. Verify those unknowns in the real Hopper workflow.
BREAKING CHANGE: Provider analysis outputs must explicitly supply metadata previously repaired with defaults; unsupported nullable facets must be reported as null.
Intern shared provenance in a graph-scoped context registry while retaining exact locations on every fact. Semantic traces carry their referenced context subset, and portable graph validation rejects stale or dangling commitments.
BREAKING CHANGE: Semantic graph facts now store evidence context IDs and locations. Graph and semantic trace results require an inline evidence_contexts table.
relationshipSignature rebuilt a node map and scanned every graph edge
for each side of each matched pair, so comparison cost grew as
pairs x (nodes + edges). Build node and adjacency indexes once per
graph in ItemContext; signatures, sorting and output are unchanged.
At 4000 matched modules, end-to-end compare: 5.7s -> 2.9s (2 edges
per node) and 10.4s -> 5.8s (6 edges per node); 8000: 3.3x and 2.5x.
Output JSON is byte-identical on the base commit.
Use the equivalent hex NUL spelling in named public contracts while preserving current path and reserved-name validation. Verify converted JSON Schema and real stdio catalog consumers.
Closes#1329
Co-authored-by: Yi-111-a <97448247+Yi-111-a@users.noreply.github.com>
Verify selected files through a nonblocking opened handle before reading. Preserve regular-file symlinks and return actionable CLI/MCP diagnostics without publishing output.
Co-authored-by: YouWish <29498139+Y0UW1SH@users.noreply.github.com>
morluto