ata-validator 1.15.0 → 1.16.0

package/index.js

@@ -3,6 +3,7 @@
 // Loading is delegated to lib/native-load.js so this file stays free of
 // platform probing and `path` (the browser entry must not pull those in).
 const native = require("./lib/native-load")();
+const { normalizeKeywords, schemaUsesKeywords } = require('./lib/keywords');
 const {
   compileToJS,
   compileToJSCodegen,
@@ -790,6 +791,16 @@ class Validator {
     this._schemaIsCallers = schemaObj === schema;
     this._vocabulariesApplied = false;
 
+    // Custom keywords, normalized once. `_usesKeywords` is what routes the
+    // schema to the interpreted engine and keeps it out of the shared
+    // compile cache; a schema that registers keywords but uses none of them
+    // takes the ordinary path.
+    this._keywords = normalizeKeywords(options.keywords);
+    this._usesKeywords = false;
+    if (this._keywords !== null && schemaUsesKeywords(schemaObj, this._keywords)) {
+      this._usesKeywords = true;
+    }
+
     this._schemaStr = null; // lazy: computed on first use
     this._schemaObj = schemaObj;
     this._options = options;
@@ -918,7 +929,7 @@ class Validator {
     const mapKey = compileCacheKey(this._schemaStr, this._schemaMap);
     // Custom formats are JS functions: bypass the compile cache since they can
     // differ between validators that share the same schema string.
-    const cached = this._userFormats ? null : _compileCache.get(mapKey);
+    const cached = (this._userFormats || this._usesKeywords) ? null : _compileCache.get(mapKey);
     let jsFn, jsCombinedFn, jsErrFn, _isCodegen = false;
     var _forceNapi = typeof process !== 'undefined' && process.env && process.env.ATA_FORCE_NAPI;
     // v1 removes the bookending requirement for $dynamicRef. Only the
@@ -936,7 +947,7 @@ class Validator {
     // survives the block and would quietly handle schemas it gets wrong; the
     // interpreted engine is both eval-free and more correct, so go straight
     // there.
-    if (this._v1Dynamic || !codegenAvailable()) {
+    if (this._v1Dynamic || this._usesKeywords || !codegenAvailable()) {
       jsFn = null; jsCombinedFn = null; jsErrFn = null;
     } else if (cached && !_forceNapi) {
       jsFn = cached.jsFn;
@@ -1030,6 +1041,7 @@ class Validator {
             schemaMap: this._schemaMap.size > 0 ? this._schemaMap : null,
             formats: this._userFormats,
             v1: isV1Dialect(schemaObj),
+            keywords: this._keywords,
           });
         }
         const r = _interp.validate(d);
@@ -1117,6 +1129,7 @@ class Validator {
             schemaMap: this._schemaMap.size > 0 ? this._schemaMap : null,
             formats: this._userFormats,
             v1: isV1Dialect(schemaObj),
+            keywords: this._keywords,
           });
         }
         const interp = _interp;
@@ -1323,6 +1336,7 @@ class Validator {
           schemaMap: this._schemaMap.size > 0 ? this._schemaMap : null,
           formats: this._userFormats,
           v1: isV1Dialect(schemaObj),
+          keywords: this._keywords,
         });
         this._engine = 'interpreter';
         _validate = (data) => interp.validate(data);
@@ -1385,6 +1399,7 @@ class Validator {
         schemaMap: this._schemaMap.size > 0 ? this._schemaMap : null,
         formats: this._userFormats,
         v1: isV1Dialect(schemaObj),
+        keywords: this._keywords,
       });
       this._engine = 'interpreter';
       if (!preprocess) this._fastVerdict = (d) => interp.isValid(d);
@@ -1632,7 +1647,7 @@ class Validator {
     // every buffer entry point goes through validate() instead.
     if (native) {
       const { bufferNeedsSlowPath, installSlowBufferApis } = require('./lib/buffer-gate.js');
-      if (bufferNeedsSlowPath(schemaObj, this._schemaMap)) installSlowBufferApis(this);
+      if (bufferNeedsSlowPath(schemaObj, this._schemaMap, this._keywords)) installSlowBufferApis(this);
     }
 
     // Save to identity cache for ultra-fast reuse with same schema object
@@ -1697,7 +1712,7 @@ class Validator {
     // A validator that rewrites its input cannot use the binding below: that
     // one answers from the compiled function alone and would skip the rewrite,
     // so isValidObject() and validate() would disagree.
-    if (this._needsPreprocess()) {
+    if (this._needsPreprocess() || this._usesKeywords) {
       this._ensureCompiled();
       return;
     }
@@ -1708,7 +1723,7 @@ class Validator {
     const mapKey = compileCacheKey(this._schemaStr, this._schemaMap);
     // Custom formats are JS functions: skip the shared cache so different
     // validators with the same schema string but different formats don't collide.
-    const cached = this._userFormats ? null : _compileCache.get(mapKey);
+    const cached = (this._userFormats || this._usesKeywords) ? null : _compileCache.get(mapKey);
     if (cached && cached.jsFn) {
       this._jsFn = cached.jsFn;
       this.isValidObject = cached.jsFn;
@@ -2131,8 +2146,11 @@ _defineLazyMethod('validate', (self) => (data) => {
 });
 _defineLazyMethod('isValidObject', (self) => (data) => {
   // A validator that rewrites its input goes through the full compile, which
-  // binds a verdict method that runs the rewrite first.
-  if (self._needsPreprocess()) {
+  // binds a verdict method that runs the rewrite first. So does one whose
+  // schema uses a custom keyword: neither the tier-0 plan nor the code
+  // generator knows the keyword, and either would accept what validate()
+  // rejects.
+  if (self._needsPreprocess() || self._usesKeywords) {
     self._ensureCompiled();
     return self.isValidObject(data);
   }

package/lib/aot-impl.js

@@ -58,6 +58,9 @@ function safeRePrelude(...fns) {
 // Generate a JS module string that can be written to a file.
 // On next startup, load with Validator.fromStandalone() -- zero compile time.
 function toStandalone(validator) {
+  if (validator._usesKeywords) {
+    throw new Error('toStandalone: a schema that uses custom keywords cannot be compiled into a standalone module (option "keywords")');
+  }
   validator._ensureCompiled();
   const jsFn = validator._jsFn;
   if (!jsFn || !jsFn._source) return null;
@@ -219,6 +222,9 @@ function emitClone(node, access, depth) {
 }
 
 function toStandaloneModule(validator, opts) {
+  if (validator._usesKeywords) {
+    throw new Error('toStandaloneModule: a schema that uses custom keywords cannot be compiled into a standalone module (option "keywords")');
+  }
   validator._ensureCompiled();
   const jsFn = validator._jsFn;
   if (!jsFn || !jsFn._source) return null;
@@ -351,6 +357,9 @@ ${parseCore}${exports}`;
 //   // On startup:
 //   const validators = Validator.loadBundle(require('./validators.js'), [schema1, schema2, ...]);
 function bundle(Validator, schemas, opts) {
+  if (opts && opts.keywords && Object.keys(opts.keywords).length > 0) {
+    throw new Error('bundle: custom keywords cannot be compiled into a standalone module (option "keywords")');
+  }
   const parts = schemas.map((schema) => {
     const v = new Validator(schema, opts);
     const standalone = toStandalone(v);
@@ -370,6 +379,12 @@ function bundle(Validator, schemas, opts) {
 // opts.format: 'cjs' (default) or 'esm'.
 // opts.formats: { name: fn } — embedded in the output via Function#toString.
 function bundleStandalone(Validator, schemas, opts) {
+  // A custom keyword is a function, and a standalone module imports nothing,
+  // so there is no way to carry it. Refuse rather than emit a module that
+  // silently ignores the keyword.
+  if (opts && opts.keywords && Object.keys(opts.keywords).length > 0) {
+    throw new Error('bundleStandalone: custom keywords cannot be compiled into a standalone module (option "keywords")');
+  }
   // Cross-schema $ref resolution: only meaningful when at least one schema has
   // an $id. Skip the schemas-as-map plumbing when none of them do.
   const haveIds = schemas.some((s) => s && typeof s === 'object' && s.$id);
@@ -436,6 +451,9 @@ function bundleStandalone(Validator, schemas, opts) {
 // Much smaller file → faster V8 parse → faster startup.
 // opts.format: 'cjs' (default) or 'esm'.
 function bundleCompact(Validator, schemas, opts) {
+  if (opts && opts.keywords && Object.keys(opts.keywords).length > 0) {
+    throw new Error('bundleCompact: custom keywords cannot be compiled into a standalone module (option "keywords")');
+  }
   const haveIds = schemas.some((s) => s && typeof s === 'object' && s.$id);
   const bundleOpts = haveIds ? { ...(opts || {}), schemas } : (opts || {});
   const format = (opts && opts.format) || 'cjs';

package/lib/buffer-gate.js

@@ -1,5 +1,7 @@
 'use strict';
 
+const { schemaUsesKeywords } = require('./keywords');
+
 // The buffer APIs (isValid, countValid, batchIsValid, isValidNDJSON,
 // isValidParallel, isValidPrepadded, and isValidJSON above the simdjson
 // threshold) answer from the native engine's own walker, which is not the
@@ -77,7 +79,11 @@ function walk(schema, depth) {
 }
 
 // True when the buffer APIs must answer through validate() for this schema.
-function bufferNeedsSlowPath(schema, schemaMap) {
+// `keywords` is the validator's normalized custom keyword map, if any: the
+// native walker knows nothing about those, so a schema using one is routed
+// through validate() like the other shapes here.
+function bufferNeedsSlowPath(schema, schemaMap, keywords) {
+  if (keywords && schemaUsesKeywords(schema, keywords)) return true;
   if (walk(schema, 0)) return true;
   if (schemaMap && schemaMap.size > 0) {
     for (const s of schemaMap.values()) if (walk(s, 1)) return true;

package/lib/compat-errors.js

@@ -0,0 +1,343 @@
+'use strict';
+
+// Shapes an error list the way the reference validator reports it, for the
+// compat entry. Two things differ between what validate() returns and what
+// `require('ajv')` users read off `validate.errors`:
+//
+// Order. The reference does not follow schema declaration order; it runs
+// keywords in a fixed order per node (the type check, the type-less
+// keywords, then the number, string, array and object groups, each in the
+// order in REFERENCE_ORDER), descending into a subschema when it reaches its
+// applicator. Consumers that show only the first error see that order.
+//
+// Wrapper errors. Where a branch or a subschema fails, the reference reports
+// the branch's own errors and then one for the keyword that combined them:
+// every failing anyOf/oneOf branch, each failing property name under
+// propertyNames, each item under contains (allErrors only), and the `if`
+// keyword after a failed then/else (allErrors only). validate() reports the
+// combining keyword alone. The shaper re-validates the branch or the value
+// against the subschema to recover those errors, so nothing here guesses.
+//
+// With allErrors off the reference stops at the first failing keyword, but
+// that keyword's wrapper group is reported whole: the branch errors and the
+// anyOf, or the property name's error and the propertyNames. The same
+// grouping is applied here after shaping.
+
+const REFERENCE_ORDER = [
+  'type',
+  '$dynamicAnchor', '$dynamicRef', '$recursiveAnchor', '$recursiveRef', '$comment', 'id', '$ref',
+  'nullable', 'const', 'enum', 'not', 'anyOf', 'oneOf', 'allOf', 'if', 'then', 'else',
+  'maximum', 'minimum', 'exclusiveMaximum', 'exclusiveMinimum', 'multipleOf',
+  'maxLength', 'minLength', 'pattern', 'format',
+  'maxItems', 'minItems', 'additionalItems', 'prefixItems', 'items', 'contains', 'uniqueItems', 'maxContains', 'minContains', 'unevaluatedItems',
+  'maxProperties', 'minProperties', 'required', 'propertyNames', 'additionalProperties', 'dependencies', 'dependentRequired', 'dependentSchemas', 'properties', 'patternProperties', 'unevaluatedProperties',
+];
+const REFERENCE_RANK = new Map(REFERENCE_ORDER.map((k, i) => [k, i]));
+
+// Keywords whose next path segment is a user-chosen name or an index rather
+// than a keyword.
+const NAMED_SEGMENT = new Set(['properties', 'patternProperties', 'dependencies', 'dependentSchemas', 'dependentRequired', '$defs', 'definitions', 'propertyDependencies']);
+const INDEXED_SEGMENT = new Set(['allOf', 'anyOf', 'oneOf', 'prefixItems', 'items', 'additionalItems']);
+
+function unescape(seg) {
+  return seg.replace(/~1/g, '/').replace(/~0/g, '~');
+}
+
+function pointerOf(schemaPath) {
+  const hash = schemaPath.indexOf('#');
+  return hash >= 0 ? schemaPath.slice(hash + 1) : schemaPath;
+}
+
+function segments(pointer) {
+  return pointer === '' ? [] : pointer.split('/').slice(1);
+}
+
+// Walks a pointer's segments and reports, per position, whether that segment
+// is a keyword or a name/index.
+function classify(segs) {
+  const kinds = new Array(segs.length);
+  let skip = false;
+  for (let i = 0; i < segs.length; i++) {
+    if (skip) { kinds[i] = 'name'; skip = false; continue; }
+    kinds[i] = 'kw';
+    const s = segs[i];
+    if (NAMED_SEGMENT.has(s)) skip = true;
+    else if (INDEXED_SEGMENT.has(s) && /^\d+$/.test(segs[i + 1] || '')) skip = true;
+  }
+  return kinds;
+}
+
+function compareReferenceOrder(a, b) {
+  const pa = segments(pointerOf(a.schemaPath || ''));
+  const pb = segments(pointerOf(b.schemaPath || ''));
+  const ka = classify(pa);
+  const n = Math.min(pa.length, pb.length);
+  for (let i = 0; i < n; i++) {
+    if (pa[i] === pb[i]) continue;
+    if (ka[i] !== 'kw') return 0;
+    const ra = REFERENCE_RANK.get(pa[i]);
+    const rb = REFERENCE_RANK.get(pb[i]);
+    if (ra === undefined || rb === undefined) return 0;
+    return ra - rb;
+  }
+  return 0;
+}
+
+function sortLikeReference(errors) {
+  return errors.length < 2 ? errors : errors.slice().sort(compareReferenceOrder);
+}
+
+function walkPointer(node, pointer) {
+  for (const raw of segments(pointer)) {
+    if (node === null || typeof node !== 'object') return undefined;
+    const seg = unescape(raw);
+    node = Array.isArray(node) ? node[Number(seg)] : node[seg];
+  }
+  return node;
+}
+
+function walkData(data, instancePath) {
+  return walkPointer(data, instancePath);
+}
+
+function escapePointer(s) {
+  return s.replace(/~/g, '~0').replace(/\//g, '~1');
+}
+
+// The keywords whose failure the reference reports after the errors of
+// what they combine. Custom keywords with a macro behave the same way and
+// validate() already reports them so.
+const WRAPPERS = new Set(['anyOf', 'oneOf', 'propertyNames', 'contains', 'not', 'if']);
+
+function createShaper({ Validator, rootId, rootDoc, schemas, options, allErrors, macroKeywords }) {
+  const branchCache = new Map();
+  const branchValidator = (pointer) => {
+    let v = branchCache.get(pointer);
+    if (!v) {
+      v = new Validator({ $ref: rootId + '#' + pointer }, options);
+      branchCache.set(pointer, v);
+    }
+    return v;
+  };
+  const wrappers = new Set(WRAPPERS);
+  for (const k of macroKeywords) wrappers.add(k);
+
+  // Errors whose schemaPath is already absolute (from the root), as opposed
+  // to relative to the branch validator that produced them.
+  const absolute = new WeakSet();
+
+  // Returns the shaped errors of `value` against the subschema at `pointer`
+  // (absolute), with schemaPath made absolute and instancePath rebased onto
+  // the caller's.
+  function subErrors(pointer, value, schemaPrefix, instancePrefix) {
+    const r = branchValidator(pointer).validate(value);
+    if (r.valid) return null;
+    const shaped = shape(r.errors, value, pointer);
+    return shaped.map((e) => {
+      const out = {
+        ...e,
+        schemaPath: absolute.has(e) ? e.schemaPath : schemaPrefix + pointer + pointerOf(e.schemaPath || '#'),
+        instancePath: instancePrefix + (e.instancePath || ''),
+      };
+      absolute.add(out);
+      return out;
+    });
+  }
+
+  function keywordError(keyword, instancePath, schemaPath, params, message) {
+    return { instancePath, schemaPath, keyword, params, message };
+  }
+
+  // `basePointer` is where `errors` were produced relative to the root
+  // document; a branch validator reports paths relative to its branch.
+  function shape(errors, data, basePointer) {
+    const sorted = sortLikeReference(errors);
+    const out = [];
+    const consumed = new Set();
+    let thenGroup = null; // { prefix, which } for the pending `if` wrapper
+
+    const flushIf = (e) => {
+      if (thenGroup === null) return;
+      const segs = segments(pointerOf(e ? e.schemaPath : ''));
+      const still = e && segs.length > thenGroup.depth && segs[thenGroup.depth] === thenGroup.which &&
+        segs.slice(0, thenGroup.depth).join('/') === thenGroup.prefixSegs;
+      if (still) return;
+      const ip = thenGroup.instancePath;
+      out.push(keywordError('if', ip, thenGroup.schemaPrefix + '/if', { failingKeyword: thenGroup.which }, `must match "${thenGroup.which}" schema`));
+      thenGroup = null;
+    };
+
+    for (let idx = 0; idx < sorted.length; idx++) {
+      const e = sorted[idx];
+      if (consumed.has(e)) continue;
+      const rel = pointerOf(e.schemaPath || '#');
+      const abs = basePointer + rel;
+      const segs = segments(rel);
+      const kinds = classify(segs);
+      const schemaPrefix = (e.schemaPath || '#').slice(0, (e.schemaPath || '#').indexOf('#') + 1);
+
+      if (allErrors) flushIf(e);
+
+      // The `if` wrapper: errors from a then/else subschema are followed by
+      // one for `if` (allErrors only, the reference returns before it
+      // otherwise).
+      if (allErrors && thenGroup === null) {
+        for (let i = 0; i < segs.length; i++) {
+          if (kinds[i] === 'kw' && (segs[i] === 'then' || segs[i] === 'else')) {
+            const prefixSegs = segs.slice(0, i);
+            thenGroup = {
+              depth: i, which: segs[i], prefixSegs: prefixSegs.join('/'),
+              schemaPrefix: schemaPrefix + (prefixSegs.length ? '/' + prefixSegs.join('/') : ''),
+              instancePath: e.instancePath,
+            };
+            break;
+          }
+        }
+      }
+
+      const kw = e.keyword;
+
+      // anyOf / oneOf: report each failing branch's errors first.
+      if ((kw === 'anyOf' || kw === 'oneOf') && segs[segs.length - 1] === kw) {
+        const branches = walkPointer(rootDoc, abs);
+        if (Array.isArray(branches)) {
+          const value = walkData(data, e.instancePath || '');
+          const branchErrors = [];
+          const passing = [];
+          for (let i = 0; i < branches.length; i++) {
+            const errs = subErrors(abs + '/' + i, value, schemaPrefix, e.instancePath || '');
+            if (errs === null) passing.push(i);
+            else branchErrors.push(allErrors ? errs : errs.slice(0, groupEnd(errs)));
+          }
+          if (passing.length === 0) {
+            for (const errs of branchErrors) out.push(...errs);
+            out.push(e);
+          } else if (kw === 'oneOf') {
+            out.push({ ...e, params: { passingSchemas: passing }, message: 'must match exactly one schema in oneOf' });
+          } else {
+            out.push(e);
+          }
+          continue;
+        }
+      }
+
+      // propertyNames: the inner errors validate() reports carry no property
+      // name. Re-check each name so every failing one is reported with its
+      // own errors and a propertyNames error, as the reference does.
+      const pnIndex = segs.findIndex((s, i) => kinds[i] === 'kw' && s === 'propertyNames');
+      if (pnIndex >= 0 && pnIndex < segs.length - 1) {
+        const pnPointer = '/' + segs.slice(0, pnIndex + 1).join('/');
+        const pnAbs = basePointer + pnPointer;
+        const object = walkData(data, e.instancePath || '');
+        if (object && typeof object === 'object' && !Array.isArray(object)) {
+          // Everything validate() said about this propertyNames node at this
+          // instance is replaced by the regenerated sequence.
+          for (let j = idx; j < sorted.length; j++) {
+            const o = sorted[j];
+            if ((o.instancePath || '') === (e.instancePath || '') && pointerOf(o.schemaPath || '#').startsWith(pnPointer + '/')) consumed.add(o);
+          }
+          const pnSchemaPath = schemaPrefix + pnPointer;
+          for (const name of Object.keys(object)) {
+            const errs = subErrors(pnAbs, name, schemaPrefix, e.instancePath || '');
+            if (errs === null) continue;
+            const kept = allErrors ? errs : errs.slice(0, 1);
+            for (const k of kept) out.push({ ...k, instancePath: e.instancePath || '', propertyName: name });
+            out.push(keywordError('propertyNames', e.instancePath || '', pnSchemaPath, { propertyName: name }, 'property name must be valid'));
+            if (!allErrors) break;
+          }
+          continue;
+        }
+      }
+
+      // contains (allErrors only): each item's errors, then the keyword's.
+      if (kw === 'contains' && allErrors && segs[segs.length - 1] === 'contains') {
+        const items = walkData(data, e.instancePath || '');
+        if (Array.isArray(items)) {
+          for (let i = 0; i < items.length; i++) {
+            const errs = subErrors(abs, items[i], schemaPrefix, (e.instancePath || '') + '/' + i);
+            if (errs !== null) out.push(...errs);
+          }
+          out.push(e);
+          continue;
+        }
+      }
+
+      // draft-07 `dependencies`, which validate() reads as dependentRequired
+      // and dependentSchemas: the reference names the original keyword.
+      const depIndex = segs.findIndex((s, i) => kinds[i] === 'kw' && (s === 'dependentRequired' || s === 'dependentSchemas'));
+      if (depIndex >= 0) {
+        const nodePointer = basePointer + (depIndex === 0 ? '' : '/' + segs.slice(0, depIndex).join('/'));
+        const node = walkPointer(rootDoc, nodePointer);
+        if (node && typeof node === 'object' && node.dependencies && typeof node.dependencies === 'object') {
+          const rewritten = segs.slice();
+          rewritten[depIndex] = 'dependencies';
+          const schemaPath = schemaPrefix + '/' + rewritten.join('/');
+          if (segs[depIndex] === 'dependentRequired' && kw === 'required') {
+            const missing = e.params && e.params.missingProperty;
+            const object = walkData(data, e.instancePath || '');
+            let property;
+            for (const [prop, deps] of Object.entries(node.dependencies)) {
+              if (Array.isArray(deps) && deps.includes(missing) && object && Object.hasOwn(object, prop)) { property = prop; break; }
+            }
+            const deps = property !== undefined ? node.dependencies[property] : [missing];
+            out.push({
+              ...e,
+              keyword: 'dependencies',
+              schemaPath,
+              params: { property, missingProperty: missing, depsCount: deps.length, deps: deps.join(', ') },
+              message: `must have ${deps.length === 1 ? 'property' : 'properties'} ${deps.join(', ')} when property ${property} is present`,
+            });
+          } else {
+            out.push({ ...e, schemaPath });
+          }
+          continue;
+        }
+      }
+
+      out.push(e);
+    }
+    if (allErrors) flushIf(null);
+    return out;
+  }
+
+  // Index just past the first error's wrapper group: the first error, plus
+  // the errors that share the path of the outermost wrapper reported later
+  // for it (the sibling branches and the anyOf, the name's errors and the
+  // propertyNames, a macro's schema errors and the macro).
+  function groupEnd(shaped) {
+    if (shaped.length < 2) return shaped.length;
+    const first = shaped[0];
+    const firstPath = pointerOf(first.schemaPath || '#');
+    let end = 1;
+    let wrapperPath = null;
+    for (let i = 1; i < shaped.length; i++) {
+      const e = shaped[i];
+      const p = pointerOf(e.schemaPath || '#');
+      if (!wrappers.has(e.keyword)) continue;
+      if (firstPath === p || firstPath.startsWith(p + '/')) {
+        if (wrapperPath === null || p.length < wrapperPath.length) { wrapperPath = p; end = i + 1; }
+      }
+    }
+    if (wrapperPath === null) return 1;
+    // Everything up to the wrapper that lives under its path belongs to
+    // the group; anything else reported in between is dropped, as the
+    // reference never reached it.
+    const group = [];
+    for (let i = 0; i < end; i++) {
+      const e = shaped[i];
+      const p = pointerOf(e.schemaPath || '#');
+      if (p === wrapperPath || p.startsWith(wrapperPath + '/')) group.push(e);
+    }
+    shaped.splice(0, shaped.length, ...group, ...shaped.slice(end));
+    return group.length;
+  }
+
+  return function shapeErrors(errors, data) {
+    const shaped = shape(errors, data, '');
+    if (allErrors) return shaped;
+    const n = groupEnd(shaped);
+    return shaped.slice(0, n);
+  };
+}
+
+module.exports = { createShaper, sortLikeReference, REFERENCE_ORDER };