From d51b7e082b73db745a1f8321e65b086902d80d80 Mon Sep 17 00:00:00 2001
From: Raymond Hill <rhill@raymondhill.net>
Date: Thu, 17 Nov 2022 08:24:55 -0500
Subject: [PATCH] Improve extraction of tokens from regexes

Fixed flawed extraction of tokens with optional sequences, i.e.
when quantifier could be zero.
Related issue:
- https://github.com/uBlockOrigin/uBlock-issues/issues/2367

Ignore look-around sequences as suggested when normalizing into
tokenizable string.
Related issue:
- https://github.com/uBlockOrigin/uBlock-issues/issues/2368

Fix regex analyzer throwing with trailing `-` in character
class sequence.
Related issue:
- https://github.com/AdguardTeam/AdguardFilters/pull/134630
---
 src/js/static-filtering-parser.js | 44 ++++++++++++++++++++++++-------
 src/lib/regexanalyzer/CHANGES.md  |  0
 src/lib/regexanalyzer/regex.js    |  6 +++++
 3 files changed, 41 insertions(+), 9 deletions(-)
 create mode 100644 src/lib/regexanalyzer/CHANGES.md

diff --git a/src/js/static-filtering-parser.js b/src/js/static-filtering-parser.js
index e62d198d8..56bc7e21b 100644
--- a/src/js/static-filtering-parser.js
+++ b/src/js/static-filtering-parser.js
@@ -3009,11 +3009,11 @@ Parser.utils = Parser.prototype.utils = (( ) => {
 
     class regex {
         static firstCharCodeClass(s) {
-            return /^[\x01%0-9A-Za-z]/.test(s) ? 1 : 0;
+            return /^[\x01\x03%0-9A-Za-z]/.test(s) ? 1 : 0;
         }
 
         static lastCharCodeClass(s) {
-            return /[\x01%0-9A-Za-z]$/.test(s) ? 1 : 0;
+            return /[\x01\x03%0-9A-Za-z]$/.test(s) ? 1 : 0;
         }
 
         static tokenizableStrFromNode(node) {
@@ -3042,18 +3042,24 @@ Parser.utils = Parser.prototype.utils = (( ) => {
                 return String.fromCharCode(firstChar, lastChar);
             }
             case 4: /* T_GROUP, 'Group' */ {
-                if ( node.flags.NegativeLookAhead === 1 ) { return '\x01'; }
-                if ( node.flags.NegativeLookBehind === 1 ) { return '\x01'; }
+                if (
+                    node.flags.LookAhead === 1 ||
+                    node.flags.NegativeLookAhead === 1 ||
+                    node.flags.LookBehind === 1 ||
+                    node.flags.NegativeLookBehind === 1
+                ) {
+                    return '';
+                }
                 return this.tokenizableStrFromNode(node.val);
             }
             case 16: /* T_QUANTIFIER, 'Quantifier' */ {
                 const s = this.tokenizableStrFromNode(node.val);
                 const first = this.firstCharCodeClass(s);
                 const last = this.lastCharCodeClass(s);
-                if ( node.flags.min === 0 && first === 0 && last === 0 ) {
-                    return '';
+                if ( node.flags.min !== 0 ) {
+                    return String.fromCharCode(first, last);
                 }
-                return String.fromCharCode(first, last);
+                return String.fromCharCode(first+2, last+2);
             }
             case 64: /* T_HEXCHAR, 'HexChar' */ {
                 return String.fromCharCode(parseInt(node.val.slice(1), 16));
@@ -3142,13 +3148,33 @@ Parser.utils = Parser.prototype.utils = (( ) => {
 
         static toTokenizableStr(reStr) {
             if ( regexAnalyzer === null ) { return ''; }
+            let s = '';
             try {
-                return this.tokenizableStrFromNode(
+                s = this.tokenizableStrFromNode(
                     regexAnalyzer(reStr, false).tree()
                 );
             } catch(ex) {
             }
-            return '';
+            // Process optional sequences
+            const reOptional = /[\x02\x03]+/g;
+            for (;;) {
+                const match = reOptional.exec(s);
+                if ( match === null ) { break; }
+                const left = s.slice(0, match.index);
+                const middle = match[0];
+                const right = s.slice(match.index + middle.length);
+                s = left;
+                s += this.firstCharCodeClass(right) === 1 ||
+                        this.firstCharCodeClass(middle) === 1
+                    ? '\x01'
+                    : '\x00';
+                s += this.lastCharCodeClass(left) === 1 ||
+                        this.lastCharCodeClass(middle) === 1
+                    ? '\x01'
+                    : '\x00';
+                s += right;
+            }
+            return s;
         }
     }
 
diff --git a/src/lib/regexanalyzer/CHANGES.md b/src/lib/regexanalyzer/CHANGES.md
new file mode 100644
index 000000000..e69de29bb
diff --git a/src/lib/regexanalyzer/regex.js b/src/lib/regexanalyzer/regex.js
index 0d206f2c7..5931d20fd 100644
--- a/src/lib/regexanalyzer/regex.js
+++ b/src/lib/regexanalyzer/regex.js
@@ -1019,6 +1019,12 @@ var rnd = function( a, b ){ return Math.round((b-a)*Math.random()+a); },
                 }
             }
 
+            if ( isRange && escaped === false && ']' === ch )
+            {
+                isRange = false;
+                chars.push('-');
+            }
+
             if ( isRange )
             {
                 if ( chars.length )