adjust regexp libraries to how unpaired surrogate are parsed now

Author: erik-krogh

javascript/ql/lib/semmle/javascript/Regexp.qll

@@ -258,8 +258,8 @@ class RegExpConstant extends RegExpTerm, @regexp_constant {
 class RegExpCharEscape extends RegExpEscape, RegExpConstant, @regexp_char_escape {
   override predicate isCharacter() {
     not (
-      // unencodable characters are represented as '?' in the database
-      getValue() = "?" and
+      // unencodable characters are represented as '?' or \uFFFD in the database
+      getValue() = ["?", 65533.toUnicode()] and
       exists(string s | s = toString().toLowerCase() |
         // only Unicode escapes give rise to unencodable characters
         s.matches("\\\\u%") and

javascript/ql/test/query-tests/Performance/ReDoS/ReDoS.expected

@@ -145,8 +145,6 @@
 | tst.js:257:14:257:116 | (.thisisagoddamnlongstringforstresstestingthequery\|\\sthisisagoddamnlongstringforstresstestingthequery)* | This part of the regular expression may cause exponential backtracking on strings containing many repetitions of ' thisisagoddamnlongstringforstresstestingthequery'. |
 | tst.js:260:14:260:77 | (thisisagoddamnlongstringforstresstestingthequery\|this\\w+query)* | This part of the regular expression may cause exponential backtracking on strings containing many repetitions of 'thisisagoddamnlongstringforstresstestingthequery'. |
 | tst.js:260:68:260:70 | \\w+ | This part of the regular expression may cause exponential backtracking on strings starting with 'this' and containing many repetitions of 'aquerythis'. |
-| tst.js:266:18:266:49 | ([\\uDC66\\uDC67]\|[\\uDC68\\uDC69])* | This part of the regular expression may cause exponential backtracking on strings starting with 'foo' and containing many repetitions of '\ufffd'. |
-| tst.js:269:18:269:51 | ((\\uDC66\|\\uDC67)\|(\\uDC68\|\\uDC69))* | This part of the regular expression may cause exponential backtracking on strings starting with 'foo' and containing many repetitions of '\ufffd'. |
 | tst.js:272:21:272:22 | b+ | This part of the regular expression may cause exponential backtracking on strings containing many repetitions of 'b'. |
 | tst.js:275:38:275:40 | \\s* | This part of the regular expression may cause exponential backtracking on strings starting with '<a a=' and containing many repetitions of '"" a='. |
 | tst.js:281:16:281:17 | a+ | This part of the regular expression may cause exponential backtracking on strings containing many repetitions of 'a'. |

javascript/ql/test/query-tests/Performance/ReDoS/tst.js

@@ -262,10 +262,10 @@ var bad61 = /(thisisagoddamnlongstringforstresstestingthequery|this\w+query)*-/
 // GOOD
 var good27 = /(thisisagoddamnlongstringforstresstestingthequery|imanotherbutunrelatedstringcomparedtotheotherstring)*-/
 
-// GOOD (but false positive caused by the extractor converting all four unpaired surrogates to \uFFFD)
+// GOOD
 var good28 = /foo([\uDC66\uDC67]|[\uDC68\uDC69])*foo/
 
-// GOOD (but false positive caused by the extractor converting all four unpaired surrogates to \uFFFD)
+// GOOD
 var good29 = /foo((\uDC66|\uDC67)|(\uDC68|\uDC69))*foo/
 
 // NOT GOOD (but cannot currently construct a prefix)

javascript/ql/test/query-tests/RegExp/DuplicateCharacterInCharacterClass/DuplicateCharacterInCharacterClass.expected

@@ -1,5 +1,4 @@
 | tst.js:1:4:1:4 | o | Character 'o' is repeated $@ in the same character class. | tst.js:1:5:1:5 | o | here |
-| tst.js:3:3:3:8 | \\uDC3A | Character '\\uDC3A' is repeated $@ in the same character class. | tst.js:3:9:3:14 | \\uDC3C | here |
 | tst.js:4:3:4:3 | ? | Character '?' is repeated $@ in the same character class. | tst.js:4:4:4:4 | ? | here |
 | tst.js:5:3:5:8 | \\u003F | Character '\\u003F' is repeated $@ in the same character class. | tst.js:5:9:5:14 | \\u003f | here |
 | tst.js:6:3:6:8 | \\u003F | Character '\\u003F' is repeated $@ in the same character class. | tst.js:6:9:6:9 | ? | here |

javascript/ql/test/query-tests/RegExp/DuplicateCharacterInCharacterClass/tst.js

@@ -1,6 +1,6 @@
 /[foo]/;
 /[a-zc]/;
-/[\uDC3A\uDC3C]/; // False positive caused by the extractor converting both unpaired surrogates to \uFFFD
+/[\uDC3A\uDC3C]/;
 /[??]/;
 /[\u003F\u003f]/;
 /[\u003F?]/;