Interpret complete scalar source at the XML owner instead of accepting numeric prefixes, permissive date strings or nonempty booleans. Retain malformed content as explicit unknown scalars with original literals and reasons, while preserving the surrounding archive and independent numeric facts.
Share the representation across plist, keyed archive, Interface Builder and entitlement projection; remove the superseded numeric fallback and helpers. Parse date fields as UTC, retain real entity/CDATA semantics, and distinguish native integer/real whitespace without ambiguous whitespace backtracking.
Validation: 276 source tests and 60 compiled/public tests passed, with updated scalar regressions rerun; static, architecture and docs checks passed. Native macOS plutil controls and CLI/MCP source-hash/time-zone controls executed. No Linux/Windows or deep-provider qualification. Tiny completed CLI task costs recorded; no performance improvement claimed.
BREAKING CHANGE: Unresolved XML scalar values now use $plist_type unknown wrappers with element, literal and reason. This replaces the former real/null unknown marker and malformed scalar prefix fallback. Supported extended UTC years now preserve their timestamp.
Fixes#4852
Interpret the validated, depth-bounded DOM directly. Collect source number
facts during the same traversal and share them with keyed archives, Interface
Builder and native plist projections. Remove the second XML parse, key aliases,
source-location rewriting and both duplicate XML numeric scanners. Preserve
character references, container/null behavior, safe own-property construction
and prototype omission accounting.
Decode signed hexadecimal integers and combine real text/CDATA. Retain exact
integer identity when a finite integer exceeds even Number's finite range,
using the shared scalar projector. Keep unsupported mixed integer syntax
conservative and report incomplete metadata.
Adapt @apitoagents's hexadecimal regression and literal-parsing proposal from
PR #3475. Native plutil controls establish the supported hex/real cases and
confirm that the synthetic 401-digit overflow case exceeds native bounds;
retaining accepted input does not claim native-format support.
Validation: 274 focused checks in 30 files pass, one skipped; build,
check:fast and docs:check pass. Built CLI/actual stdio MCP, native plist and
Interface Builder controls cover kinds, references, exact digits and digests.
Eight baseline checks fail; native binary conversion independently verifies
updated CDATA real expectations. 48 tiny completed CLI cost comparisons
preserve ordinary XML/binary and all three native replies byte for byte.
Two read-only reviews clear. No statistical speed claim, heavy profiler,
investigated-target execution or real deep-provider verification.
Fixes#3373Fixes#4042Fixes#4789
Use the validated DOM's character-data key names for both identity and
prototype omission. Replace first-child reinterpretation with canonical,
collision-free aliases inside the pinned decoder boundary; restore names
through the existing dictionary projection. Alias every spelling of the
same name to preserve duplicate assignment order. Handle empty key spellings
and keep scalar key elements outside aliasing. Preserve value text and bytes.
Adapt @thangldw's tests and proposal from PR #3426. Avoid the proposal's
closing-content scanner and re-escaping source names. Add duplicate, empty,
alias-collision, scalar-key, native fallback and Interface Builder controls.
Validation: 180 focused checks across 20 files pass, including built CLI and
actual stdio MCP, build, check:fast and docs:check. Baseline public inspection
truncates main to ma. 24 tiny completed CLI cost controls preserve ordinary
XML/binary full replies and Evidence IDs byte for byte. No statistical speed
claim, heavy profiling, investigated-target or real deep-provider execution.
Fixes#3371
Reuse byte-derived numeric metadata after graph classification so emitted
attributes preserve exact integers without changing UID links or identities.
Reserve metadata and attribute projection within the shared decode budget.
Validated six CLI cases covering XML, binary, collisions, real values and
incomplete metadata, plus decode-budget regressions. A small cold/warm CLI
comparison showed similar costs; it does not establish a speed improvement.
Refs #3876, #3972.
Co-authored-by: hhang <169027046+xu91102@users.noreply.github.com>
morluto