[webkit-changes] [274806] trunk

Mon, 22 Mar 2021 15:08:03 -0700

Title: [274806] trunk
Revision
274806
Author
[email protected]
Date
2021-03-22 15:06:52 -0700 (Mon, 22 Mar 2021)

Log Message

[YARR] Interpreter incorrectly matches non-BMP characters with multiple .
https://bugs.webkit.org/show_bug.cgi?id=223498

Reviewed by Yusuke Suzuki.

JSTests:

New test.

* stress/regexp-dot-match-nonBMP.js: Added.
(shouldMatch):
(shouldntMatch):

Source/_javascript_Core:

We need to check that we read an actual character before seeing if it is part of a character class.
In the case where we are checking that a character is not in a character class, like .,
the failed to read result from input.readChecked(), -1, is not part of the newline character class.
This will allow regular expressions that require more than the number of characters in a string
to match.

* yarr/YarrInterpreter.cpp:
(JSC::Yarr::Interpreter::checkCharacterClass):

LayoutTests:

Updated test.

* fast/forms/ValidityState-patternMismatch-expected.txt:
* fast/forms/ValidityState-patternMismatch.html:

Modified Paths

Added Paths

Diff

Modified: trunk/JSTests/ChangeLog (274805 => 274806)


--- trunk/JSTests/ChangeLog	2021-03-22 22:05:48 UTC (rev 274805)
+++ trunk/JSTests/ChangeLog	2021-03-22 22:06:52 UTC (rev 274806)
@@ -1,3 +1,16 @@
+2021-03-22  Michael Saboff  <[email protected]>
+
+        [YARR] Interpreter incorrectly matches non-BMP characters with multiple .
+        https://bugs.webkit.org/show_bug.cgi?id=223498
+
+        Reviewed by Yusuke Suzuki.
+
+        New test.
+
+        * stress/regexp-dot-match-nonBMP.js: Added.
+        (shouldMatch):
+        (shouldntMatch):
+
 2021-03-22  Yusuke Suzuki  <[email protected]>
 
         [JSC] Intl.Locale should not assume is8Bit

Added: trunk/JSTests/stress/regexp-dot-match-nonBMP.js (0 => 274806)


--- trunk/JSTests/stress/regexp-dot-match-nonBMP.js	                        (rev 0)
+++ trunk/JSTests/stress/regexp-dot-match-nonBMP.js	2021-03-22 22:06:52 UTC (rev 274806)
@@ -0,0 +1,20 @@
+function shouldMatch(regexp, str) {
+    let result = regexp.test(str);
+    if (result !== true)
+        throw new Error("Expected " + regexp + ".test(\"" + str + "\") to be true, but wasn't");
+}
+
+function shouldntMatch(regexp, str) {
+    let result = regexp.test(str);
+    if (result !== false)
+        throw new Error("Expected " + regexp + ".test(\"" + str + "\") to be false, but wasn't");
+}
+
+let s = String.fromCodePoint(0x10000);
+
+shouldMatch(/./, s);
+shouldMatch(/./u, s);
+shouldMatch(/../, s);
+shouldntMatch(/../u, s);
+shouldntMatch(/.../, s);
+shouldntMatch(/.../u, s);

Modified: trunk/LayoutTests/ChangeLog (274805 => 274806)


--- trunk/LayoutTests/ChangeLog	2021-03-22 22:05:48 UTC (rev 274805)
+++ trunk/LayoutTests/ChangeLog	2021-03-22 22:06:52 UTC (rev 274806)
@@ -1,3 +1,15 @@
+2021-03-22  Michael Saboff  <[email protected]>
+
+        [YARR] Interpreter incorrectly matches non-BMP characters with multiple .
+        https://bugs.webkit.org/show_bug.cgi?id=223498
+
+        Reviewed by Yusuke Suzuki.
+
+        Updated test.
+
+        * fast/forms/ValidityState-patternMismatch-expected.txt:
+        * fast/forms/ValidityState-patternMismatch.html:
+
 2021-03-22  Chris Gambrell  <[email protected]>
 
         [LayoutTests] Convert http/tests/multipart convert PHP to Python

Modified: trunk/LayoutTests/fast/forms/ValidityState-patternMismatch-expected.txt (274805 => 274806)


--- trunk/LayoutTests/fast/forms/ValidityState-patternMismatch-expected.txt	2021-03-22 22:05:48 UTC (rev 274805)
+++ trunk/LayoutTests/fast/forms/ValidityState-patternMismatch-expected.txt	2021-03-22 22:06:52 UTC (rev 274806)
@@ -29,7 +29,6 @@
 PASS patternMismatchFor("match-19") is false
 PASS patternMismatchFor("match-20") is false
 PASS patternMismatchFor("match-21") is false
-PASS patternMismatchFor("match-22") is false
 PASS patternMismatchFor("wrong-gray-or-grey") is true
 PASS patternMismatchFor("gray") is false
 PASS patternMismatchFor("grey") is false
@@ -56,6 +55,7 @@
 PASS patternMismatchFor("mismatch-20") is true
 PASS patternMismatchFor("mismatch-21") is true
 PASS patternMismatchFor("mismatch-22") is true
+PASS patternMismatchFor("mismatch-23") is true
 PASS patternMismatchFor("empty-pattern-match") is false
 PASS patternMismatchFor("empty-pattern-mismatch") is true
 PASS patternMismatchFor("invalid-01") is false

Modified: trunk/LayoutTests/fast/forms/ValidityState-patternMismatch.html (274805 => 274806)


--- trunk/LayoutTests/fast/forms/ValidityState-patternMismatch.html	2021-03-22 22:05:48 UTC (rev 274805)
+++ trunk/LayoutTests/fast/forms/ValidityState-patternMismatch.html	2021-03-22 22:06:52 UTC (rev 274806)
@@ -33,7 +33,6 @@
 <input id="match-19" type="text" pattern="10|11|12|[0-9]" value="12" />
 <input id="match-20" type="text" pattern="f(o|e)\1" value="foo" />
 <input id="match-21" type="text" pattern="a.b" value="a&#x1D306;b" />
-<input id="match-22" type="text" pattern="a..b" value="a&#x1D306;b" />
 <input id="empty-pattern-match" type="text" pattern="" value="" />
 <input id="wrong-gray-or-grey" type="text" pattern="gr[ae]y" value="Wrong!"
 /><input id="gray" type="text" pattern="gr[ae]y" value="gray"
@@ -60,7 +59,8 @@
 /><input id="mismatch-19" type="text" pattern="^" value="wrong"
 /><input id="mismatch-20" type="text" pattern="$" value="wrong"
 /><input id="mismatch-21" type="text" pattern="f(o|e)\1" value="foe"
-/><input id="mismatch-22" type="text" pattern="a...b" value="a&#x1D306;b" />
+/><input id="mismatch-22" type="text" pattern="a..b" value="a&#x1D306;b" />
+<input id="mismatch-23" type="text" pattern="a...b" value="a&#x1D306;b" />
 
 <input id="invalid-01" type="text" pattern=")foo(" value="foo"/>
 <input id="invalid-02" type="text" pattern=")foo(" value="foobar"/>
@@ -107,7 +107,6 @@
 shouldBeFalse('patternMismatchFor("match-19")');
 shouldBeFalse('patternMismatchFor("match-20")');
 shouldBeFalse('patternMismatchFor("match-21")');
-shouldBeFalse('patternMismatchFor("match-22")');
 
 shouldBeTrue('patternMismatchFor("wrong-gray-or-grey")');
 shouldBeFalse('patternMismatchFor("gray")');
@@ -136,6 +135,7 @@
 shouldBeTrue('patternMismatchFor("mismatch-20")');
 shouldBeTrue('patternMismatchFor("mismatch-21")');
 shouldBeTrue('patternMismatchFor("mismatch-22")');
+shouldBeTrue('patternMismatchFor("mismatch-23")');
 
 shouldBeFalse('patternMismatchFor("empty-pattern-match")');
 shouldBeTrue('patternMismatchFor("empty-pattern-mismatch")');

Modified: trunk/Source/_javascript_Core/ChangeLog (274805 => 274806)


--- trunk/Source/_javascript_Core/ChangeLog	2021-03-22 22:05:48 UTC (rev 274805)
+++ trunk/Source/_javascript_Core/ChangeLog	2021-03-22 22:06:52 UTC (rev 274806)
@@ -1,3 +1,19 @@
+2021-03-22  Michael Saboff  <[email protected]>
+
+        [YARR] Interpreter incorrectly matches non-BMP characters with multiple .
+        https://bugs.webkit.org/show_bug.cgi?id=223498
+
+        Reviewed by Yusuke Suzuki.
+
+        We need to check that we read an actual character before seeing if it is part of a character class.
+        In the case where we are checking that a character is not in a character class, like .,
+        the failed to read result from input.readChecked(), -1, is not part of the newline character class.
+        This will allow regular expressions that require more than the number of characters in a string
+        to match.
+
+        * yarr/YarrInterpreter.cpp:
+        (JSC::Yarr::Interpreter::checkCharacterClass):
+
 2021-03-22  Ross Kirsling  <[email protected]>
 
         Unreviewed, fix Mac and non-unified JSCOnly builds

Modified: trunk/Source/_javascript_Core/yarr/YarrInterpreter.cpp (274805 => 274806)


--- trunk/Source/_javascript_Core/yarr/YarrInterpreter.cpp	2021-03-22 22:05:48 UTC (rev 274805)
+++ trunk/Source/_javascript_Core/yarr/YarrInterpreter.cpp	2021-03-22 22:06:52 UTC (rev 274806)
@@ -425,7 +425,11 @@
 
     bool checkCharacterClass(CharacterClass* characterClass, bool invert, unsigned negativeInputOffset)
     {
-        bool match = testCharacterClass(characterClass, input.readChecked(negativeInputOffset));
+        int inputChar = input.readChecked(negativeInputOffset);
+        if (inputChar < 0)
+            return false;
+
+        bool match = testCharacterClass(characterClass, inputChar);
         return invert ? !match : match;
     }
     

_______________________________________________
webkit-changes mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-changes

Reply via email to