eyeling 1.14.13 → 1.14.14

package/eyeling.js

@@ -7847,6 +7847,95 @@ function decodeN3StringEscapes(s, offset = null) {
   return out;
 }
 
+function formatCodePoint(cp) {
+  return cp
+    .toString(16)
+    .toUpperCase()
+    .padStart(cp <= 0xffff ? 4 : 6, '0');
+}
+
+function isForbiddenIriRefChar(c) {
+  return (
+    c === '<' || c === '>' || c === '"' || c === '{' || c === '}' || c === '|' || c === '^' || c === '`' || c === '\\'
+  );
+}
+
+function assertValidIriRefCodePoint(cp, offset = null) {
+  if (cp <= 0x20) {
+    throw new N3SyntaxError(`Invalid IRIREF: character U+${formatCodePoint(cp)} is not allowed inside <...>`, offset);
+  }
+
+  if (cp >= 0xd800 && cp <= 0xdfff) {
+    throw new N3SyntaxError(
+      `Invalid IRIREF: surrogate code point U+${formatCodePoint(cp)} is not allowed inside <...>`,
+      offset,
+    );
+  }
+
+  if (isForbiddenNoncharacterCodePoint(cp)) {
+    throw new N3SyntaxError(
+      `Invalid IRIREF: noncharacter U+${formatCodePoint(cp)} is not allowed inside <...>`,
+      offset,
+    );
+  }
+
+  const c = String.fromCodePoint(cp);
+  if (isForbiddenIriRefChar(c)) {
+    throw new N3SyntaxError(`Invalid IRIREF: character ${JSON.stringify(c)} is not allowed inside <...>`, offset);
+  }
+}
+
+function decodeIriRefEscapes(s, offset = null) {
+  let out = '';
+  for (let i = 0; i < s.length; i++) {
+    const c = s[i];
+    if (c !== '\\') {
+      const cp = c.codePointAt(0);
+      assertValidIriRefCodePoint(cp, offset);
+      out += c;
+      continue;
+    }
+
+    if (i + 1 >= s.length) {
+      throw new N3SyntaxError('Invalid IRIREF: bare backslash is not allowed inside <...>', offset);
+    }
+
+    const e = s[++i];
+    if (e === 'u') {
+      const hex = s.slice(i + 1, i + 5);
+      if (!/^[0-9A-Fa-f]{4}$/.test(hex)) {
+        throw new N3SyntaxError('Invalid IRIREF: malformed \\u escape inside <...>', offset);
+      }
+      const cp = parseInt(hex, 16);
+      assertValidIriRefCodePoint(cp, offset);
+      out += String.fromCodePoint(cp);
+      i += 4;
+      continue;
+    }
+
+    if (e === 'U') {
+      const hex = s.slice(i + 1, i + 9);
+      if (!/^[0-9A-Fa-f]{8}$/.test(hex)) {
+        throw new N3SyntaxError('Invalid IRIREF: malformed \\U escape inside <...>', offset);
+      }
+      const cp = parseInt(hex, 16);
+      if (cp < 0 || cp > 0x10ffff) {
+        throw new N3SyntaxError(`Invalid IRIREF: code point U+${hex.toUpperCase()} is out of range`, offset);
+      }
+      assertValidIriRefCodePoint(cp, offset);
+      out += String.fromCodePoint(cp);
+      i += 8;
+      continue;
+    }
+
+    throw new N3SyntaxError(
+      `Invalid IRIREF: character ${JSON.stringify('\\' + e)} is not allowed inside <...>`,
+      offset,
+    );
+  }
+  return out;
+}
+
 function assertValidStringLiteralValue(s, offset = null) {
   for (let i = 0; i < s.length; i++) {
     const cu = s.charCodeAt(i);
@@ -8032,7 +8121,7 @@ function lex(inputText) {
         throw new N3SyntaxError('Unterminated IRI <...>', start);
       }
       i++; // skip '>'
-      const iri = iriChars.join('');
+      const iri = decodeIriRefEscapes(iriChars.join(''), start);
       tokens.push(new Token('IriRef', iri, start));
       continue;
     }

package/lib/lexer.js

@@ -201,6 +201,95 @@ function decodeN3StringEscapes(s, offset = null) {
   return out;
 }
 
+function formatCodePoint(cp) {
+  return cp
+    .toString(16)
+    .toUpperCase()
+    .padStart(cp <= 0xffff ? 4 : 6, '0');
+}
+
+function isForbiddenIriRefChar(c) {
+  return (
+    c === '<' || c === '>' || c === '"' || c === '{' || c === '}' || c === '|' || c === '^' || c === '`' || c === '\\'
+  );
+}
+
+function assertValidIriRefCodePoint(cp, offset = null) {
+  if (cp <= 0x20) {
+    throw new N3SyntaxError(`Invalid IRIREF: character U+${formatCodePoint(cp)} is not allowed inside <...>`, offset);
+  }
+
+  if (cp >= 0xd800 && cp <= 0xdfff) {
+    throw new N3SyntaxError(
+      `Invalid IRIREF: surrogate code point U+${formatCodePoint(cp)} is not allowed inside <...>`,
+      offset,
+    );
+  }
+
+  if (isForbiddenNoncharacterCodePoint(cp)) {
+    throw new N3SyntaxError(
+      `Invalid IRIREF: noncharacter U+${formatCodePoint(cp)} is not allowed inside <...>`,
+      offset,
+    );
+  }
+
+  const c = String.fromCodePoint(cp);
+  if (isForbiddenIriRefChar(c)) {
+    throw new N3SyntaxError(`Invalid IRIREF: character ${JSON.stringify(c)} is not allowed inside <...>`, offset);
+  }
+}
+
+function decodeIriRefEscapes(s, offset = null) {
+  let out = '';
+  for (let i = 0; i < s.length; i++) {
+    const c = s[i];
+    if (c !== '\\') {
+      const cp = c.codePointAt(0);
+      assertValidIriRefCodePoint(cp, offset);
+      out += c;
+      continue;
+    }
+
+    if (i + 1 >= s.length) {
+      throw new N3SyntaxError('Invalid IRIREF: bare backslash is not allowed inside <...>', offset);
+    }
+
+    const e = s[++i];
+    if (e === 'u') {
+      const hex = s.slice(i + 1, i + 5);
+      if (!/^[0-9A-Fa-f]{4}$/.test(hex)) {
+        throw new N3SyntaxError('Invalid IRIREF: malformed \\u escape inside <...>', offset);
+      }
+      const cp = parseInt(hex, 16);
+      assertValidIriRefCodePoint(cp, offset);
+      out += String.fromCodePoint(cp);
+      i += 4;
+      continue;
+    }
+
+    if (e === 'U') {
+      const hex = s.slice(i + 1, i + 9);
+      if (!/^[0-9A-Fa-f]{8}$/.test(hex)) {
+        throw new N3SyntaxError('Invalid IRIREF: malformed \\U escape inside <...>', offset);
+      }
+      const cp = parseInt(hex, 16);
+      if (cp < 0 || cp > 0x10ffff) {
+        throw new N3SyntaxError(`Invalid IRIREF: code point U+${hex.toUpperCase()} is out of range`, offset);
+      }
+      assertValidIriRefCodePoint(cp, offset);
+      out += String.fromCodePoint(cp);
+      i += 8;
+      continue;
+    }
+
+    throw new N3SyntaxError(
+      `Invalid IRIREF: character ${JSON.stringify('\\' + e)} is not allowed inside <...>`,
+      offset,
+    );
+  }
+  return out;
+}
+
 function assertValidStringLiteralValue(s, offset = null) {
   for (let i = 0; i < s.length; i++) {
     const cu = s.charCodeAt(i);
@@ -386,7 +475,7 @@ function lex(inputText) {
         throw new N3SyntaxError('Unterminated IRI <...>', start);
       }
       i++; // skip '>'
-      const iri = iriChars.join('');
+      const iri = decodeIriRefEscapes(iriChars.join(''), start);
       tokens.push(new Token('IriRef', iri, start));
       continue;
     }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "eyeling",
-  "version": "1.14.13",
+  "version": "1.14.14",
   "description": "A minimal Notation3 (N3) reasoner in JavaScript.",
   "main": "./index.js",
   "keywords": [

package/test/api.test.js

@@ -567,6 +567,227 @@ bad.:example a bad.:Person.
 `,
     expectError: true,
   },
+  {
+    name: '12h invalid syntax: space is not allowed inside IRIREF',
+    opt: { proofComments: false },
+    input: `
+@prefix : <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+<http://bad example.org/> a :BadExample.
+
+{
+    :subject :predicate ?X.
+}
+=>
+{
+    :result :has :crash-syntax-7.
+}.
+
+{} => {
+    :test :contains :crash-syntax-7.
+}.
+
+{
+    :result :has :crash-syntax-7.
+}
+=>
+{
+    :test :is false.
+}.
+`,
+    expectError: true,
+  },
+  {
+    name: '12i invalid syntax: UCHAR escape is not allowed inside IRIREF',
+    opt: { proofComments: false },
+    input: String.raw`
+@prefix : <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+<http://bad\u0020example.org/> a :BadExample.
+
+{
+    :subject :predicate ?X.
+}
+=>
+{
+    :result :has :crash-syntax-8.
+}.
+
+{} => {
+    :test :contains :crash-syntax-8.
+}.
+
+{
+    :result :has :crash-syntax-8.
+}
+=>
+{
+    :test :is false.
+}.
+`,
+    expectError: true,
+  },
+  {
+    name: '12j invalid syntax: control characters are not allowed inside IRIREF',
+    opt: { proofComments: false },
+    input: `
+@prefix : <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+<http://badexample.org> :foo <http://example.org/> .
+
+{
+    :subject :predicate ?X.
+}
+=>
+{
+    :result :has :crash-syntax-9.
+}.
+
+{} => {
+    :test :contains :crash-syntax-9.
+}.
+
+{
+    :result :has :crash-syntax-9.
+}
+=>
+{
+    :test :is false.
+}.
+`,
+    expectError: true,
+  },
+  {
+    name: '12k invalid syntax: control-character UCHAR is not allowed inside IRIREF',
+    opt: { proofComments: false },
+    input: String.raw`
+@prefix : <http://example.org/> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+
+<http://bad\u0007example.org> :foo <http://example.org/> .
+
+{
+    :subject :predicate ?X.
+}
+=>
+{
+    :result :has :crash-syntax-10.
+}.
+
+{} => {
+    :test :contains :crash-syntax-10.
+}.
+
+{
+    :result :has :crash-syntax-10.
+}
+=>
+{
+    :test :is false.
+}.
+`,
+    expectError: true,
+  },
+  {
+    name: '12l regression: IRIREF \\u escape decodes before log:uri comparison (mismatch stays falsey)',
+    opt: { proofComments: false },
+    input: String.raw`
+@prefix : <http://example.org/> .
+@prefix log: <http://www.w3.org/2000/10/swap/log#> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@base <http://example.org/>.
+
+{
+    <http://example.org/\u0041> log:uri "http://example.org/\\u0041".
+}
+=>
+{
+    :result :has :fail-literal-1.
+}.
+
+{ } => {
+    :test :contains :fail-literal-1.
+}.
+`,
+    expect: [/:(?:test)\s+:(?:contains)\s+:(?:fail-literal-1)\s*\./],
+    notExpect: [/:(?:result)\s+:(?:has)\s+:(?:fail-literal-1)\s*\./, /:(?:test)\s+:(?:is)\s+true\s*\./],
+  },
+  {
+    name: '12m regression: IRIREF \\u escape matches plain-A literal via log:uri',
+    opt: { proofComments: false },
+    input: String.raw`
+@prefix : <http://example.org/> .
+@prefix log: <http://www.w3.org/2000/10/swap/log#> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@base <http://example.org/>.
+
+{
+    <http://example.org/\u0041> log:uri "http://example.org/A".
+}
+=>
+{
+    :result :has :success-literal-5.
+}.
+
+{ } => {
+    :test :contains :success-literal-5.
+}.
+
+{
+    :result :has :success-literal-5.
+}
+=>
+{
+    :test :is true.
+}.
+`,
+    expect: [
+      /:(?:result)\s+:(?:has)\s+:(?:success-literal-5)\s*\./,
+      /:(?:test)\s+:(?:contains)\s+:(?:success-literal-5)\s*\./,
+      /:(?:test)\s+:(?:is)\s+true\s*\./,
+    ],
+  },
+  {
+    name: '12n regression: IRIREF \\u escape matches decoded literal escape via log:uri',
+    opt: { proofComments: false },
+    input: String.raw`
+@prefix : <http://example.org/> .
+@prefix log: <http://www.w3.org/2000/10/swap/log#> .
+@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
+@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
+@base <http://example.org/>.
+
+{
+    <http://example.org/\u0041> log:uri "http://example.org/\u0041".
+}
+=>
+{
+    :result :has :success-literal-6.
+}.
+
+{ } => {
+    :test :contains :success-literal-6.
+}.
+
+{
+    :result :has :success-literal-6.
+}
+=>
+{
+    :test :is true.
+}.
+`,
+    expect: [
+      /:(?:result)\s+:(?:has)\s+:(?:success-literal-6)\s*\./,
+      /:(?:test)\s+:(?:contains)\s+:(?:success-literal-6)\s*\./,
+      /:(?:test)\s+:(?:is)\s+true\s*\./,
+    ],
+  },
   {
     name: '13 heavier recursion: ancestor closure over 15 links',
     opt: { proofComments: false, maxBuffer: 200 * 1024 * 1024 },