@adguard/agtree 2.1.3 → 2.2.0

package/dist/agtree.d.ts

@@ -1,5 +1,5 @@
 /*
- * AGTree v2.1.3 (build date: Mon, 21 Oct 2024 08:22:25 GMT)
+ * AGTree v2.2.0 (build date: Wed, 27 Nov 2024 16:28:27 GMT)
  * (c) 2024 Adguard Software Ltd.
  * Released under the MIT license
  * https://github.com/AdguardTeam/tsurlfilter/tree/master/packages/agtree#readme
@@ -2357,9 +2357,11 @@ declare class FilterListParser extends ParserBase {
      * @param ast AST to generate
      * @param preferRaw If `true`, then the parser will use `raws.text` property of each rule
      * if it is available. Default is `false`.
+     * @param tolerant If `true`, errors during rule generation will be logged to the console and invalid rules
+     * will be skipped. If `false`, an error will be thrown on the first invalid rule. Default is `true`.
      * @returns Serialized filter list
      */
-    static generate(ast: FilterList, preferRaw?: boolean): string;
+    static generate(ast: FilterList, preferRaw?: boolean, tolerant?: boolean): string;
     /**
      * Serializes a filter list node to binary format.
      *
@@ -3697,6 +3699,7 @@ declare const modifierDataSchema: zod.ZodEffects<zod.ZodTypeAny, {
     exceptionOnly: boolean;
     valueOptional: boolean;
     valueFormat: string | null;
+    valueFormatFlags: string | null;
 }, any>;
 /**
  * Type of the modifier data schema.

package/dist/agtree.js

@@ -1,5 +1,5 @@
 /*
- * AGTree v2.1.3 (build date: Mon, 21 Oct 2024 08:22:25 GMT)
+ * AGTree v2.2.0 (build date: Wed, 27 Nov 2024 16:28:27 GMT)
  * (c) 2024 Adguard Software Ltd.
  * Released under the MIT license
  * https://github.com/AdguardTeam/tsurlfilter/tree/master/packages/agtree#readme
@@ -13,7 +13,6 @@ var sprintfJs = require('sprintf-js');
 var cssTokenizer = require('@adguard/css-tokenizer');
 var tldts = require('tldts');
 var isIp = require('is-ip');
-var XRegExp = require('xregexp');
 var cloneDeep = require('clone-deep');
 var compatibilityTableData_js = require('./compatibility-table-data.js');
 var zod = require('zod');
@@ -1913,6 +1912,9 @@ class AgentCommentRuleParser extends ParserBase {
     static serialize(node, buffer) {
         buffer.writeUint8(BinaryTypeMap.AgentRuleNode);
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             buffer.writeUint8(AgentRuleSerializationMap.Children);
             // note: we store the count, because re-construction of the array is faster if we know the length
@@ -1967,6 +1969,11 @@ class AgentCommentRuleParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
 }
 
@@ -2862,6 +2869,9 @@ class HintCommentRuleParser extends ParserBase {
             buffer.writeUint8(SYNTAX_SERIALIZATION_MAP.get(exports.AdblockSyntax.Adg) ?? 0);
         }
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             buffer.writeUint8(HintRuleSerializationMap.Children);
             // note: we store the count, because re-construction of the array is faster if we know the length
@@ -2915,6 +2925,11 @@ class HintCommentRuleParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
 }
 
@@ -5348,6 +5363,9 @@ class ModifierListParser extends ParserBase {
     static serialize(node, buffer) {
         buffer.writeUint8(BinaryTypeMap.ModifierListNode);
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             buffer.writeUint8(ModifierListNodeSerializationMap.Children);
             // note: we store the count, because re-construction of the array is faster if we know the length
@@ -5399,6 +5417,11 @@ class ModifierListParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
 }
 
@@ -8611,10 +8634,16 @@ class NetworkRuleParser extends ParserBase {
         const pattern = ValueParser.parse(raw.slice(patternStart, patternEnd), options, baseOffset + patternStart);
         // Parse modifiers (if any)
         let modifiers;
+        // Get a last non-whitespace index
+        const lastNonWsIndex = StringUtils.skipWSBack(raw);
         // Find start and end index of the modifiers
         const modifiersStart = separatorIndex + 1;
-        const modifiersEnd = StringUtils.skipWSBack(raw) + 1;
+        const modifiersEnd = lastNonWsIndex + 1;
         if (separatorIndex !== -1) {
+            // Check for empty modifiers
+            if (separatorIndex === lastNonWsIndex) {
+                throw new AdblockSyntaxError('Empty modifiers are not allowed', baseOffset + separatorIndex, baseOffset + raw.length);
+            }
             modifiers = ModifierListParser.parse(raw.slice(modifiersStart, modifiersEnd), options, baseOffset + modifiersStart);
         }
         // Throw error if there is no pattern and no modifiers
@@ -8933,6 +8962,9 @@ class HostRuleParser extends ParserBase {
             buffer.writeUint32(node.end);
         }
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             // note: we store the count, because re-construction of the array is faster if we know the length
             if (count > UINT16_MAX) {
@@ -8976,6 +9008,11 @@ class HostRuleParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
     /**
      * Serializes a host rule node to binary format.
@@ -9743,9 +9780,11 @@ class FilterListParser extends ParserBase {
      * @param ast AST to generate
      * @param preferRaw If `true`, then the parser will use `raws.text` property of each rule
      * if it is available. Default is `false`.
+     * @param tolerant If `true`, errors during rule generation will be logged to the console and invalid rules
+     * will be skipped. If `false`, an error will be thrown on the first invalid rule. Default is `true`.
      * @returns Serialized filter list
      */
-    static generate(ast, preferRaw = false) {
+    static generate(ast, preferRaw = false, tolerant = true) {
         let result = EMPTY;
         for (let i = 0; i < ast.children.length; i += 1) {
             const rule = ast.children[i];
@@ -9753,7 +9792,18 @@ class FilterListParser extends ParserBase {
                 result += rule.raws.text;
             }
             else {
-                result += RuleParser.generate(rule);
+                try {
+                    result += RuleParser.generate(rule);
+                }
+                catch (error) {
+                    if (tolerant) {
+                        // eslint-disable-next-line no-console
+                        console.error(`Error when generating: ${error}`);
+                    }
+                    else {
+                        throw new Error(String(error));
+                    }
+                }
             }
             switch (rule.raws?.nl) {
                 case 'crlf':
@@ -10713,10 +10763,11 @@ const isCustomValueFormatValidator = (valueFormat) => {
  *
  * @param modifier Modifier AST node.
  * @param valueFormat Value format for the modifier.
+ * @param valueFormatFlags Optional; RegExp flags for the value format.
  *
  * @returns Validation result.
  */
-const validateValue = (modifier, valueFormat) => {
+const validateValue = (modifier, valueFormat, valueFormatFlags) => {
     if (isCustomValueFormatValidator(valueFormat)) {
         const validator = CUSTOM_VALUE_FORMAT_MAP[valueFormat];
         return validator(modifier);
@@ -10725,14 +10776,19 @@ const validateValue = (modifier, valueFormat) => {
     if (!modifier.value?.value) {
         return getValueRequiredValidationResult(modifierName);
     }
-    let xRegExp;
+    let regExp;
     try {
-        xRegExp = XRegExp(valueFormat);
+        if (isString(valueFormatFlags)) {
+            regExp = new RegExp(valueFormat, valueFormatFlags);
+        }
+        else {
+            regExp = new RegExp(valueFormat);
+        }
     }
     catch (e) {
         throw new Error(`${SOURCE_DATA_ERROR_PREFIX.INVALID_VALUE_FORMAT_REGEXP}: '${modifierName}'`);
     }
-    const isValid = xRegExp.test(modifier.value?.value);
+    const isValid = regExp.test(modifier.value?.value);
     if (!isValid) {
         return getInvalidValidationResult(`${VALIDATION_ERROR_PREFIX.VALUE_INVALID}: '${modifierName}'`);
     }
@@ -11316,7 +11372,7 @@ const validateForSpecificSyntax = (syntax, modifier, isException) => {
         if (!specificBlockerData.valueFormat) {
             throw new Error(`${SOURCE_DATA_ERROR_PREFIX.NO_VALUE_FORMAT_FOR_ASSIGNABLE}: '${modifierName}'`);
         }
-        return validateValue(modifier, specificBlockerData.valueFormat);
+        return validateValue(modifier, specificBlockerData.valueFormat, specificBlockerData.valueFormatFlags);
     }
     if (modifier?.value) {
         // e.g. 'third-party=true'
@@ -13013,8 +13069,14 @@ function getErrorMessage(error) {
  * Known validators that don't need to be validated as regex.
  */
 const KNOWN_VALIDATORS = new Set([
+    'csp_value',
     'domain',
+    'permissions_value',
+    'pipe_separated_apps',
+    'pipe_separated_denyallow_domains',
     'pipe_separated_domains',
+    'pipe_separated_methods',
+    'pipe_separated_stealth_options',
     'regexp',
     'url',
 ]);
@@ -13065,6 +13127,10 @@ zodToCamelCase(baseCompatibilityDataSchema.extend({
      * Its value can be a regex pattern or a known validator name (e.g. `domain`, `pipe_separated_domains`, etc.).
      */
     value_format: nonEmptyStringSchema.nullable().default(null),
+    /**
+     * Describes the flags for the `value_format` regex pattern.
+     */
+    value_format_flags: nonEmptyStringSchema.nullable().default(null),
 }).superRefine((data, ctx) => {
     // TODO: find something better, for now we can't add refine logic to the base schema:
     // https://github.com/colinhacks/zod/issues/454#issuecomment-848370721
@@ -13085,11 +13151,17 @@ zodToCamelCase(baseCompatibilityDataSchema.extend({
         const valueFormat = data.value_format.trim();
         // if it is a known validator, we don't need to validate it further
         if (KNOWN_VALIDATORS.has(valueFormat)) {
+            if (data.value_format_flags) {
+                ctx.addIssue({
+                    code: zod.ZodIssueCode.custom,
+                    message: 'value_format_flags are not allowed for known validators',
+                });
+            }
             return;
         }
         // otherwise, we need to validate it as a regex
         try {
-            XRegExp(valueFormat);
+            new RegExp(valueFormat, data.value_format_flags ?? EMPTY);
         }
         catch (error) {
             ctx.addIssue({
@@ -13098,6 +13170,12 @@ zodToCamelCase(baseCompatibilityDataSchema.extend({
             });
         }
     }
+    else if (data.value_format_flags) {
+        ctx.addIssue({
+            code: zod.ZodIssueCode.custom,
+            message: 'value_format is required for value_format_flags',
+        });
+    }
 }));
 
 /**
@@ -14700,7 +14778,7 @@ class RawFilterListConverter extends ConverterBase {
             return createConversionResult(rawFilterList, false);
         }
         // Otherwise, serialize the filter list and return the result
-        return createConversionResult(FilterListParser.generate(conversionResult.result), true);
+        return createConversionResult(FilterListParser.generate(conversionResult.result, false, tolerant), true);
     }
 }
 
@@ -15648,7 +15726,7 @@ class RuleCategorizer {
     }
 }
 
-const version = "2.1.3";
+const version = "2.2.0";
 
 /**
  * @file AGTree version

package/dist/agtree.mjs

@@ -1,5 +1,5 @@
 /*
- * AGTree v2.1.3 (build date: Mon, 21 Oct 2024 08:22:25 GMT)
+ * AGTree v2.2.0 (build date: Wed, 27 Nov 2024 16:28:27 GMT)
  * (c) 2024 Adguard Software Ltd.
  * Released under the MIT license
  * https://github.com/AdguardTeam/tsurlfilter/tree/master/packages/agtree#readme
@@ -12,7 +12,6 @@ import { TokenType as TokenType$1, tokenizeExtended, getFormattedTokenName } fro
 import * as tldts from 'tldts';
 import { parse } from 'tldts';
 import isIp from 'is-ip';
-import XRegExp from 'xregexp';
 import cloneDeep from 'clone-deep';
 import { modifiersCompatibilityTableData, redirectsCompatibilityTableData, scriptletsCompatibilityTableData } from './compatibility-table-data.js';
 import zod from 'zod';
@@ -1893,6 +1892,9 @@ class AgentCommentRuleParser extends ParserBase {
     static serialize(node, buffer) {
         buffer.writeUint8(BinaryTypeMap.AgentRuleNode);
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             buffer.writeUint8(AgentRuleSerializationMap.Children);
             // note: we store the count, because re-construction of the array is faster if we know the length
@@ -1947,6 +1949,11 @@ class AgentCommentRuleParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
 }
 
@@ -2842,6 +2849,9 @@ class HintCommentRuleParser extends ParserBase {
             buffer.writeUint8(SYNTAX_SERIALIZATION_MAP.get(AdblockSyntax.Adg) ?? 0);
         }
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             buffer.writeUint8(HintRuleSerializationMap.Children);
             // note: we store the count, because re-construction of the array is faster if we know the length
@@ -2895,6 +2905,11 @@ class HintCommentRuleParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
 }
 
@@ -5328,6 +5343,9 @@ class ModifierListParser extends ParserBase {
     static serialize(node, buffer) {
         buffer.writeUint8(BinaryTypeMap.ModifierListNode);
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             buffer.writeUint8(ModifierListNodeSerializationMap.Children);
             // note: we store the count, because re-construction of the array is faster if we know the length
@@ -5379,6 +5397,11 @@ class ModifierListParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
 }
 
@@ -8591,10 +8614,16 @@ class NetworkRuleParser extends ParserBase {
         const pattern = ValueParser.parse(raw.slice(patternStart, patternEnd), options, baseOffset + patternStart);
         // Parse modifiers (if any)
         let modifiers;
+        // Get a last non-whitespace index
+        const lastNonWsIndex = StringUtils.skipWSBack(raw);
         // Find start and end index of the modifiers
         const modifiersStart = separatorIndex + 1;
-        const modifiersEnd = StringUtils.skipWSBack(raw) + 1;
+        const modifiersEnd = lastNonWsIndex + 1;
         if (separatorIndex !== -1) {
+            // Check for empty modifiers
+            if (separatorIndex === lastNonWsIndex) {
+                throw new AdblockSyntaxError('Empty modifiers are not allowed', baseOffset + separatorIndex, baseOffset + raw.length);
+            }
             modifiers = ModifierListParser.parse(raw.slice(modifiersStart, modifiersEnd), options, baseOffset + modifiersStart);
         }
         // Throw error if there is no pattern and no modifiers
@@ -8913,6 +8942,9 @@ class HostRuleParser extends ParserBase {
             buffer.writeUint32(node.end);
         }
         const count = node.children.length;
+        // If there are no children, we do not write any data related to them, to avoid using unnecessary storage,
+        // but children is a required field, so during deserialization we should initialize it as an empty array,
+        // if there are no children in the binary data.
         if (count) {
             // note: we store the count, because re-construction of the array is faster if we know the length
             if (count > UINT16_MAX) {
@@ -8956,6 +8988,11 @@ class HostRuleParser extends ParserBase {
             }
             prop = buffer.readUint8();
         }
+        // Maybe children are not present in the binary data,
+        // in this case, we should initialize it as an empty array.
+        if (!node.children) {
+            node.children = [];
+        }
     }
     /**
      * Serializes a host rule node to binary format.
@@ -9723,9 +9760,11 @@ class FilterListParser extends ParserBase {
      * @param ast AST to generate
      * @param preferRaw If `true`, then the parser will use `raws.text` property of each rule
      * if it is available. Default is `false`.
+     * @param tolerant If `true`, errors during rule generation will be logged to the console and invalid rules
+     * will be skipped. If `false`, an error will be thrown on the first invalid rule. Default is `true`.
      * @returns Serialized filter list
      */
-    static generate(ast, preferRaw = false) {
+    static generate(ast, preferRaw = false, tolerant = true) {
         let result = EMPTY;
         for (let i = 0; i < ast.children.length; i += 1) {
             const rule = ast.children[i];
@@ -9733,7 +9772,18 @@ class FilterListParser extends ParserBase {
                 result += rule.raws.text;
             }
             else {
-                result += RuleParser.generate(rule);
+                try {
+                    result += RuleParser.generate(rule);
+                }
+                catch (error) {
+                    if (tolerant) {
+                        // eslint-disable-next-line no-console
+                        console.error(`Error when generating: ${error}`);
+                    }
+                    else {
+                        throw new Error(String(error));
+                    }
+                }
             }
             switch (rule.raws?.nl) {
                 case 'crlf':
@@ -10693,10 +10743,11 @@ const isCustomValueFormatValidator = (valueFormat) => {
  *
  * @param modifier Modifier AST node.
  * @param valueFormat Value format for the modifier.
+ * @param valueFormatFlags Optional; RegExp flags for the value format.
  *
  * @returns Validation result.
  */
-const validateValue = (modifier, valueFormat) => {
+const validateValue = (modifier, valueFormat, valueFormatFlags) => {
     if (isCustomValueFormatValidator(valueFormat)) {
         const validator = CUSTOM_VALUE_FORMAT_MAP[valueFormat];
         return validator(modifier);
@@ -10705,14 +10756,19 @@ const validateValue = (modifier, valueFormat) => {
     if (!modifier.value?.value) {
         return getValueRequiredValidationResult(modifierName);
     }
-    let xRegExp;
+    let regExp;
     try {
-        xRegExp = XRegExp(valueFormat);
+        if (isString(valueFormatFlags)) {
+            regExp = new RegExp(valueFormat, valueFormatFlags);
+        }
+        else {
+            regExp = new RegExp(valueFormat);
+        }
     }
     catch (e) {
         throw new Error(`${SOURCE_DATA_ERROR_PREFIX.INVALID_VALUE_FORMAT_REGEXP}: '${modifierName}'`);
     }
-    const isValid = xRegExp.test(modifier.value?.value);
+    const isValid = regExp.test(modifier.value?.value);
     if (!isValid) {
         return getInvalidValidationResult(`${VALIDATION_ERROR_PREFIX.VALUE_INVALID}: '${modifierName}'`);
     }
@@ -11296,7 +11352,7 @@ const validateForSpecificSyntax = (syntax, modifier, isException) => {
         if (!specificBlockerData.valueFormat) {
             throw new Error(`${SOURCE_DATA_ERROR_PREFIX.NO_VALUE_FORMAT_FOR_ASSIGNABLE}: '${modifierName}'`);
         }
-        return validateValue(modifier, specificBlockerData.valueFormat);
+        return validateValue(modifier, specificBlockerData.valueFormat, specificBlockerData.valueFormatFlags);
     }
     if (modifier?.value) {
         // e.g. 'third-party=true'
@@ -12993,8 +13049,14 @@ function getErrorMessage(error) {
  * Known validators that don't need to be validated as regex.
  */
 const KNOWN_VALIDATORS = new Set([
+    'csp_value',
     'domain',
+    'permissions_value',
+    'pipe_separated_apps',
+    'pipe_separated_denyallow_domains',
     'pipe_separated_domains',
+    'pipe_separated_methods',
+    'pipe_separated_stealth_options',
     'regexp',
     'url',
 ]);
@@ -13045,6 +13107,10 @@ zodToCamelCase(baseCompatibilityDataSchema.extend({
      * Its value can be a regex pattern or a known validator name (e.g. `domain`, `pipe_separated_domains`, etc.).
      */
     value_format: nonEmptyStringSchema.nullable().default(null),
+    /**
+     * Describes the flags for the `value_format` regex pattern.
+     */
+    value_format_flags: nonEmptyStringSchema.nullable().default(null),
 }).superRefine((data, ctx) => {
     // TODO: find something better, for now we can't add refine logic to the base schema:
     // https://github.com/colinhacks/zod/issues/454#issuecomment-848370721
@@ -13065,11 +13131,17 @@ zodToCamelCase(baseCompatibilityDataSchema.extend({
         const valueFormat = data.value_format.trim();
         // if it is a known validator, we don't need to validate it further
         if (KNOWN_VALIDATORS.has(valueFormat)) {
+            if (data.value_format_flags) {
+                ctx.addIssue({
+                    code: zod.ZodIssueCode.custom,
+                    message: 'value_format_flags are not allowed for known validators',
+                });
+            }
             return;
         }
         // otherwise, we need to validate it as a regex
         try {
-            XRegExp(valueFormat);
+            new RegExp(valueFormat, data.value_format_flags ?? EMPTY);
         }
         catch (error) {
             ctx.addIssue({
@@ -13078,6 +13150,12 @@ zodToCamelCase(baseCompatibilityDataSchema.extend({
             });
         }
     }
+    else if (data.value_format_flags) {
+        ctx.addIssue({
+            code: zod.ZodIssueCode.custom,
+            message: 'value_format is required for value_format_flags',
+        });
+    }
 }));
 
 /**
@@ -14680,7 +14758,7 @@ class RawFilterListConverter extends ConverterBase {
             return createConversionResult(rawFilterList, false);
         }
         // Otherwise, serialize the filter list and return the result
-        return createConversionResult(FilterListParser.generate(conversionResult.result), true);
+        return createConversionResult(FilterListParser.generate(conversionResult.result, false, tolerant), true);
     }
 }
 
@@ -15628,7 +15706,7 @@ class RuleCategorizer {
     }
 }
 
-const version = "2.1.3";
+const version = "2.2.0";
 
 /**
  * @file AGTree version