Changeset 73594 in webkit for trunk/JavaScriptCore/yarr/RegexParser.h

Timestamp:

Dec 8, 2010, 9:40:29 PM (15 years ago)

Author:

[email protected]

Message:

Permit Character Class Escape in CharacterRange in Character Class.
​https://bugs.webkit.org/show_bug.cgi?id=50483
​https://bugs.webkit.org/show_bug.cgi?id=50538
​https://bugs.webkit.org/show_bug.cgi?id=50654
​https://bugs.webkit.org/show_bug.cgi?id=50646

Reviewed by Sam Weinig.

We recently tightened up our spec conformance in generating syntax
error in these cases, however testing in the wild has shown this
to be problematic. This reverts the previous change in allowing
class escapes (e.g. \d) in ranges in character classes ([]), but
does retain some closer conformance to the spec in only allowing
ranges that would be permitted per the grammar rules in the spec
(e.g. in /[\d-a-z]/ "a-z" cannot be considered as a range).

JavaScriptCore:

• yarr/RegexParser.h:

(JSC::Yarr::Parser::CharacterClassParserDelegate::atomPatternCharacter):
(JSC::Yarr::Parser::CharacterClassParserDelegate::atomBuiltInCharacterClass):
(JSC::Yarr::Parser::parse):

LayoutTests:

• fast/js/regexp-ranges-and-escaped-hyphens-expected.txt:
• fast/js/script-tests/regexp-ranges-and-escaped-hyphens.js:
• fast/regex/invalid-range-in-class-expected.txt:
• fast/regex/pcre-test-1-expected.txt:
• fast/regex/script-tests/invalid-range-in-class.js:
• fast/regex/script-tests/pcre-test-1.js:

File:

• trunk/JavaScriptCore/yarr/RegexParser.h (modified) (4 diffs)

trunk/JavaScriptCore/yarr/RegexParser.h

@@ -59 +59 @@
         ParenthesesTypeInvalid,
         CharacterClassUnmatched,
-        CharacterClassInvalidRange,
         CharacterClassOutOfOrder,
         EscapeUnterminated,
@@ -143 +142 @@
                 return;
 
+                // See coment in atomBuiltInCharacterClass below.
+                // This too is technically an error, per ECMA-262, and again we
+                // we chose to allow this.  Note a subtlely here that while we
+                // diverge from the spec's definition of CharacterRange we do
+                // remain in compliance with the grammar.  For example, consider
+                // the expression /[\d-a-z]/.  We comply with the grammar in
+                // this case by not allowing a-z to be matched as a range.
             case AfterCharacterClassHyphen:
-                // Error! We have something like /[\d-x]/.
-                m_err = CharacterClassInvalidRange;
+                m_delegate.atomCharacterClassAtom(ch);
+                m_state = Empty;
                 return;
             }
@@ -168 +174 @@
                 return;
 
+                // If we hit either of these cases, we have an invalid range that
+                // looks something like /[x-\d]/ or /[\d-\d]/.
+                // According to ECMA-262 this should be a syntax error, but
+                // empirical testing shows this to break teh webz.  Instead we
+                // comply with to the ECMA-262 grammar, and assume the grammar to
+                // have matched the range correctly, but tweak our interpretation
+                // of CharacterRange.  Effectively we implicitly handle the hyphen
+                // as if it were escaped, e.g. /[\w-_]/ is treated as /[\w\-_]/.
             case CachedCharacterHyphen:
+                m_delegate.atomCharacterClassAtom(m_character);
+                m_delegate.atomCharacterClassAtom('-');
+                // fall through
             case AfterCharacterClassHyphen:
-                // Error! If we hit either of these cases, we have an
-                // invalid range that looks something like /[x-\d]/
-                // or /[\d-\d]/.
-                m_err = CharacterClassInvalidRange;
+                m_delegate.atomCharacterClassBuiltIn(classID, invert);
+                m_state = Empty;
                 return;
             }
@@ -682 +697 @@
             "unrecognized character after (?",
             "missing terminating ] for character class",
-            "invalid range in character class",
             "range out of order in character class",
             "\\ at end of pattern"