Release Schema Tools v0.1.2

public/CHANGELOG.md

@@ -1,5 +1,17 @@
 # Changelog
 
+## 0.1.2 - 2026-09-01
+
+- Refused invalid sample claims when `false` JSON Schemas are reached through local references or required combiners, while selecting the first viable `anyOf` or `oneOf` branch.
+- Centralized sample budgets across JSON Schema, OpenAPI, XSD, Relax NG, and Schematron output.
+- Applied the 512 KiB aggregate text budget to repeated XML names, attributes, and values, and corrected XML/JSON ceilings to never exceed 2,000 generated values or elements and exactly 20 generated nesting levels.
+- Made text truncation UTF-16-safe so it never splits a surrogate pair and corrupts generated XML.
+- Preserved direct-root Relax NG `element` patterns in generated samples.
+- Added a monotonic 50,000-step work ceiling that refuses over-budget generation, so repeated references, failed alternatives, and non-emitting pattern fan-out cannot amplify work or bypass later mandatory constraints.
+- Replaced costly live DOM child-collection conversions with linear sibling-pointer walks and cached bounded XSD type shapes.
+- Cached reused schema/reference and XML pattern lookups, and omitted duplicate derived XML attributes so generated samples remain well formed.
+- Added adversarial regressions for referenced boolean schemas, exhausted combiners and required values, amplified shared XSD values, wide XML containers, repeated Relax NG references, and exact JSON/XML node limits.
+
 ## 0.1.1 - 2026-09-01
 
 - Bounded literal values, aggregate text, derived XML names, collections, and final output during sample generation.

public/README.md

@@ -14,7 +14,7 @@ Schema Tools is a standalone local-first application in the [add·ideas Toolbox]
 - Conservative comparisons for required properties, types, enums, declarations, operations, parameters, and responses
 - Inert rendering, offline PWA support, responsive Toolbox shell integration, and deterministic release archives
 
-Each document is limited to 2 MiB of text, the workspace to 8 MiB and 20 documents, and parsed trees, references, instances, samples, validation work, and diagnostic output have independent bounds. Generated samples are additionally limited to 2 MiB of serialized output, 512 KiB of retained text, and 1,024 characters per derived literal. DTD/entity declarations, remote references, Schematron XPath, extension code, JSON Schema `pattern`, and `patternProperties` are never executed. JSON Schema validation covers a documented assertion subset rather than claiming full specification conformance. XML languages receive structural inspection—not instance validation—and OpenAPI checks are intentionally focused.
+Each document is limited to 2 MiB of text, the workspace to 8 MiB and 20 documents, and parsed trees, references, instances, samples, validation work, and diagnostic output have independent bounds. One shared generator budget limits samples to 50,000 monotonic work steps, 2,000 generated JSON values or XML elements, exactly 20 generated nesting levels, 512 KiB of aggregate derived text, 1,024 UTF-16 code units per literal without splitting surrogate pairs, and 2 MiB of serialized output. Generation is refused if another work step would exceed that ceiling; structural and text ceilings instead return a visibly bounded heuristic sample where one can still be formed. DTD/entity declarations, remote references, Schematron XPath, extension code, JSON Schema `pattern`, and `patternProperties` are never executed. JSON Schema validation covers a documented assertion subset rather than claiming full specification conformance. XML languages receive structural inspection—not instance validation—and OpenAPI checks are intentionally focused.
 
 See [Architecture](docs/ARCHITECTURE.md) and [Privacy and security](docs/PRIVACY-SECURITY.md) for the exact capability boundaries.
 
@@ -30,7 +30,7 @@ npm run test:browser
 
 ## Release
 
-`npm run release:artifact` creates deterministic `release/schema-tools-0.1.1.zip` and checksum files.
+`npm run release:artifact` creates deterministic `release/schema-tools-0.1.2.zip` and checksum files.
 
 ## Licence
 

public/SOURCE.md

@@ -1,7 +1,7 @@
 # Corresponding source
 
-The corresponding source for Schema Tools 0.1.1 is available at:
+The corresponding source for Schema Tools 0.1.2 is available at:
 
-https://git.add-ideas.de/lotobo/schema-tools/src/tag/v0.1.1
+https://git.add-ideas.de/lotobo/schema-tools/src/tag/v0.1.2
 
 Build with Node.js 22, npm 11, `npm ci`, and `npm run release:artifact`.

public/THIRD_PARTY_NOTICES.md

@@ -1,6 +1,6 @@
 # Third-party notices
 
-Schema Tools 0.1.1 directly depends on these runtime packages:
+Schema Tools 0.1.2 directly depends on these runtime packages:
 
 | Package                          | Pinned version | Declared licence |
 | -------------------------------- | -------------: | ---------------- |

public/docs/ARCHITECTURE.md

@@ -6,6 +6,8 @@ JSON and YAML values are converted into an acyclic, prototype-safe JSON model wi
 
 Every reference is classified before validation. Only fragments and relative filenames supplied in the current workspace are eligible. Root `$id`/legacy `id` values are deliberately ignored so relative references remain anchored to workspace filenames; nested identifier scopes, named anchors, dynamic/recursive references, and unevaluated keywords are outside the subset and cause a visible refusal. All participating JSON Schema files must use one declared draft. Formats and unknown extension keywords are annotations. Schemas containing `pattern` or `patternProperties` are refused because JavaScript regular-expression execution cannot be reliably time-bounded; they remain inspectable. Decimal arithmetic uses JavaScript numbers, so `multipleOf` applies a small floating-point tolerance.
 
-OpenAPI JSON/YAML receives focused document, operation, response, reference, sample, and comparison logic. It does not run requests and does not claim full OpenAPI conformance. XML uses the browser's inert `DOMParser` only after rejecting DTD and entity declarations. XSD, Relax NG XML syntax, and Schematron are checked for well-formedness and structurally inventoried. Schematron XPath and extensions are retained as text and never executed. XSD and Relax NG sample generation deliberately follows a bounded first branch and is labelled heuristic. Across languages, sample generation copies literals rather than retaining source objects, caps individual derived values and names, limits aggregate retained text and nodes, and rejects serialized output above 2 MiB.
+OpenAPI JSON/YAML receives focused document, operation, response, reference, sample, and comparison logic. It does not run requests and does not claim full OpenAPI conformance. XML uses the browser's inert `DOMParser` only after rejecting DTD and entity declarations. Element counting and depth inspection use a linear sibling-pointer walk, avoiding repeated conversion of live DOM child collections. XSD, Relax NG XML syntax, and Schematron are checked for well-formedness and structurally inventoried. Schematron XPath and extensions are retained as text and never executed. XSD and Relax NG sample generation deliberately follows a bounded first branch and is labelled heuristic. XSD named-type shapes are inspected once and cached within one generation; Relax NG grammars and direct-root `element` patterns use the same renderer, and pattern-only wrappers do not consume generated nesting depth.
 
-The PWA uses only relative URLs, so the same build works standalone or below a nested portal route. Its service worker caches same-origin files from its own scope. No worker, storage database, server API, telemetry path, or remote schema resolver exists in v0.1.1.
+All sample paths share one budget implementation. It admits at most 50,000 monotonic generator work steps and 2,000 generated JSON values or XML elements across exactly 20 generated levels (the root is level one), accounts for at most 512 KiB of aggregate derived keys, names, attribute values, and content, caps each copied literal at 1,024 UTF-16 code units without splitting a surrogate pair, and rejects serialized output above 2 MiB. References and schema/pattern combiners do not consume generated depth, but every build, copy, reference, alternative, XML type inspection, and XML pattern visit consumes work before expansion. Per-generation caches ensure repeated JSON references, wide property collections, XML child collections, inline types, text, and name normalization are not rescanned without bound. The work counter is deliberately not restored when a failed heuristic alternative rolls back its node/text checkpoint, and generation is refused rather than returning an ambiguously partial result when a 50,001st work step is attempted. Node, depth, and text ceilings can instead omit bounded material with a visible notice. Repeated references consume node and text counters for every emitted occurrence. Duplicate or fallback-colliding derived XML attribute names are omitted with a notice to preserve well-formed output. A `false` JSON Schema reached through a selected local reference or mandatory `allOf` branch aborts generation; `anyOf` and `oneOf` skip definitely impossible boolean branches and use the first viable heuristic branch only while the complete attempt remains within budget.
+
+The PWA uses only relative URLs, so the same build works standalone or below a nested portal route. Its service worker caches same-origin files from its own scope. No worker, storage database, server API, telemetry path, or remote schema resolver exists in v0.1.2.

public/docs/PRIVACY-SECURITY.md

@@ -2,8 +2,8 @@
 
 Schema sources, instances, diagnostics, comparisons, and generated samples stay in page memory. There are no accounts, analytics, telemetry, persistence, remote imports, or runtime third-party assets. Explicit source links are normal navigation only. Clearing or closing the page releases application references but cannot promise forensic erasure from browser or operating-system memory.
 
-Each source is limited to 2 MiB of text; a workspace is limited to 20 documents and 8 MiB. JSON/YAML trees are capped at 25,000 values, 48 levels, and 2,000 entries in one collection. XML is capped at 25,000 elements and 48 levels. References, instances, and diagnostics have separate caps. Sample generation is limited to 2,000 nodes, 20 levels, 512 KiB of retained text, 1,024 characters per copied literal, and 2 MiB of serialized output. File byte gates are deliberately conservative before `File.text()` decoding.
+Each source is limited to 2 MiB of text; a workspace is limited to 20 documents and 8 MiB. JSON/YAML trees are capped at 25,000 values, 48 levels, and 2,000 entries in one collection. XML is capped at 25,000 elements and 48 levels. References, instances, and diagnostics have separate caps. A shared sample budget is limited to 50,000 monotonic work steps, exactly 2,000 generated JSON values or XML elements, exactly 20 generated nesting levels, 512 KiB of aggregate derived text, 1,024 UTF-16 code units per copied literal without splitting surrogate pairs, and 2 MiB of serialized output. Reusing a local definition consumes the same aggregate counters on every attempt and generated occurrence; failed heuristic alternatives never refund work. Reaching the work ceiling refuses the generation operation, so exhaustion cannot be mistaken for a viable choice or skip a later mandatory constraint. File byte gates are deliberately conservative before `File.text()` decoding.
 
 Prototype-sensitive JSON keys, cyclic YAML aliases, NUL input, DTD/entity declarations, absolute/remote/escaping references, and executable schema extensions are rejected. Schematron expressions and imported XML are never executed. JSON Schema formats are annotations, and regex-bearing schema keywords are not executed. The focused validator interprets eligible schemas without dynamic code generation, remote loading, custom code, or `unsafe-eval`; work and diagnostic counts are capped.
 
-Sample generation and compatibility results are review aids. A generated document is not guaranteed to satisfy every constraint, and an absence of reported changes does not prove compatibility. JSON validation requires a single declared draft and uses local workspace filenames rather than identifier URIs. Named anchors, nested identifier scopes, dynamic/recursive references, unevaluated keywords, regular expressions, and full meta-schema validation are outside v0.1.1. XSD, Relax NG, and Schematron instance validation is also outside v0.1.1; OpenAPI inspection is not a full conformance certification.
+Sample generation and compatibility results are review aids. A generated document is not guaranteed to satisfy every constraint, and an absence of reported changes does not prove compatibility. Generation refuses a definitely impossible selected `false` JSON Schema branch but does not prove broader satisfiability. JSON validation requires a single declared draft and uses local workspace filenames rather than identifier URIs. Named anchors, nested identifier scopes, dynamic/recursive references, unevaluated keywords, regular expressions, and full meta-schema validation are outside v0.1.2. XSD, Relax NG, and Schematron instance validation is also outside v0.1.2; OpenAPI inspection is not a full conformance certification.

public/sw.js

@@ -1,5 +1,5 @@
 const CACHE_PREFIX = "schema-tools-shell-";
-const CACHE_NAME = CACHE_PREFIX + "0.1.1";
+const CACHE_NAME = CACHE_PREFIX + "0.1.2";
 const CORE = ["./", "./manifest.webmanifest", "./favicon.svg"];
 self.addEventListener("install", (event) => {
   event.waitUntil(

public/toolbox-app.json

@@ -3,7 +3,7 @@
   "schemaVersion": 1,
   "id": "de.add-ideas.schema-tools",
   "name": "Schema Tools",
-  "version": "0.1.1",
+  "version": "0.1.2",
   "description": "Inspect, validate, compare, and derive schema examples locally.",
   "entry": "./",
   "icon": "./favicon.svg",