Bug 578714: Stop doing Unicode-aware case-folding inside layout/style. r=zwol sr=bz

--HG--
extra : rebase_source : 7d5f43bf8d7d109c9669db68589e26554df9a250

content/base/public/nsContentUtils.h

@@ -1493,12 +1493,14 @@ public:
   /**
    * Convert ASCII A-Z to a-z.
    */
+  static void ASCIIToLower(nsAString& aStr);
   static void ASCIIToLower(const nsAString& aSource, nsAString& aDest);
 
   /**
    * Convert ASCII a-z to A-Z.
    */
   static void ASCIIToUpper(nsAString& aStr);
+  static void ASCIIToUpper(const nsAString& aSource, nsAString& aDest);
 
   static nsIInterfaceRequestor* GetSameOriginChecker();
 

content/base/src/nsContentUtils.cpp

@@ -4948,6 +4948,21 @@ nsContentUtils::GetCurrentJSContext()
   return cx;
 }
 
+/* static */
+void
+nsContentUtils::ASCIIToLower(nsAString& aStr)
+{
+  PRUnichar* iter = aStr.BeginWriting();
+  PRUnichar* end = aStr.EndWriting();
+  while (iter != end) {
+    PRUnichar c = *iter;
+    if (c >= 'A' && c <= 'Z') {
+      *iter = c + ('a' - 'A');
+    }
+    ++iter;
+  }
+}
+
 /* static */
 void
 nsContentUtils::ASCIIToLower(const nsAString& aSource, nsAString& aDest)
@@ -4983,6 +4998,26 @@ nsContentUtils::ASCIIToUpper(nsAString& aStr)
   }
 }
 
+/* static */
+void
+nsContentUtils::ASCIIToUpper(const nsAString& aSource, nsAString& aDest)
+{
+  PRUint32 len = aSource.Length();
+  aDest.SetLength(len);
+  if (aDest.Length() == len) {
+    PRUnichar* dest = aDest.BeginWriting();
+    const PRUnichar* iter = aSource.BeginReading();
+    const PRUnichar* end = aSource.EndReading();
+    while (iter != end) {
+      PRUnichar c = *iter;
+      *dest = (c >= 'a' && c <= 'z') ?
+         c + ('A' - 'a') : c;
+      ++iter;
+      ++dest;
+    }
+  }
+}
+
 PRBool
 nsContentUtils::EqualsIgnoreASCIICase(const nsAString& aStr1,
                                       const nsAString& aStr2)

intl/unicharutil/util/nsUnicharUtils.cpp

@@ -186,6 +186,14 @@ ToLowerCase(const nsAString& aSource,
   ToLowerCase(in, out, len);
 }
 
+PRUnichar
+ToLowerCaseASCII(const PRUnichar aChar)
+{
+  if (IS_ASCII_UPPER(aChar))
+    return aChar + 0x0020;
+  return aChar;
+}
+
 void
 ToUpperCase(nsAString& aString)
 {
@@ -234,6 +242,49 @@ nsCaseInsensitiveStringComparator::operator()(PRUnichar lhs,
     return 1;
 }
 
+PRInt32
+nsASCIICaseInsensitiveStringComparator::operator()(const PRUnichar* lhs,
+                                                   const PRUnichar* rhs,
+                                                   PRUint32 aLength) const
+{
+  while (aLength) {
+    PRUnichar l = *lhs++;
+    PRUnichar r = *lhs++;
+    if (l != r) {
+      l = ToLowerCaseASCII(l);
+      r = ToLowerCaseASCII(r);
+      
+      if (l > r)
+        return 1;
+      else if (r > l)
+        return -1;
+    }
+    aLength--;
+  }
+
+  return 0;
+}
+
+PRInt32
+nsASCIICaseInsensitiveStringComparator::operator()(PRUnichar lhs,
+                                                   PRUnichar rhs) const
+{
+  // see if they're an exact match first
+  if (lhs == rhs)
+    return 0;
+  
+  lhs = ToLowerCaseASCII(lhs);
+  rhs = ToLowerCaseASCII(rhs);
+  
+  if (lhs == rhs)
+    return 0;
+  else if (lhs < rhs)
+    return -1;
+  else
+    return 1;
+}
+
+
 #endif // MOZILLA_INTERNAL_API
 
 PRInt32

intl/unicharutil/util/nsUnicharUtils.h

@@ -92,6 +92,16 @@ public:
   }
 };
 
+class nsASCIICaseInsensitiveStringComparator : public nsStringComparator
+{
+public:
+  virtual int operator() (const PRUnichar*,
+                          const PRUnichar*,
+                          PRUint32 aLength) const;
+  virtual int operator() (PRUnichar,
+                          PRUnichar) const;
+};
+
 inline PRBool
 CaseInsensitiveFindInReadable(const nsAString& aPattern,
                               nsAString::const_iterator& aSearchStart,

layout/reftests/reftest.list

@@ -217,6 +217,9 @@ include text-transform/reftest.list
 # -moz-transform/
 include transform/reftest.list
 
+# unicode/ (verify that we don't do expend effort doing unicode-aware case checks)
+include unicode/reftest.list
+
 # widget/
 include ../../widget/reftests/reftest.list
 

layout/reftests/unicode/reftest.list

@@ -0,0 +1,6 @@
+== unicode-attribute-selector.html unicode-ref.html
+== unicode-element-selector.html unicode-ref.html
+== unicode-lang.html unicode-ref.html
+== unicode-media-query-media-type.html unicode-ref-print.html
+== unicode-media-query-query.html unicode-ref-print.html
+== unicode-pseudo-selector.html unicode-ref.html

layout/reftests/unicode/unicode-attribute-selector.html

@@ -0,0 +1,19 @@
+<html>
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - attribute selectors</title>
+  <style>
+    div[dİr] {color: red;}
+    div[dİr="ltr"] {text-indent: 30px;}
+    input[dİsabled] {display: none;}
+    //input[dİsabled="disabled"] {display: inline;}
+  </style>
+</head>
+
+<body>
+  <div dir='ltr'><p lang="hi">स्टार</p></div>
+  <input disabled>
+  <input dİsabled="DİSABLED">
+  <input dİsabled="disabled">
+</body>
+</html>

layout/reftests/unicode/unicode-element-selector.html

@@ -0,0 +1,14 @@
+<html>
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - element selectors</title>
+  <style>
+    dİv {color: red}
+  </style>
+</head>
+
+<body>
+  <div><p lang="hi">स्टार</p></div>
+  <input disabled>
+</body>
+</html>

layout/reftests/unicode/unicode-lang.html

@@ -0,0 +1,15 @@
+<html>
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - language selector</title>
+  <style>
+    p:lang(hİ) {color: red}
+    //p[lang="hİ"] {text-indent:30px} Disabled until Bug 492431 is fixed
+  </style>
+</head>
+
+<body>
+  <div><p lang="hi">स्टार</p></div>
+  <input disabled>
+</body>
+</html>

layout/reftests/unicode/unicode-media-query-media-type.html

@@ -0,0 +1,15 @@
+<html class="reftest-print">
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - media query - media type selector</title>
+  <style>
+    @media prİnt {
+    p { color: red }
+    }
+  </style>
+</head>
+
+<body>
+  <div><p lang="hi">स्टार</p></div>
+</body>
+</html>

layout/reftests/unicode/unicode-media-query-query.html

@@ -0,0 +1,13 @@
+<html class="reftest-print">
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - media query - media query text</title>
+  <style>
+    @media print and (mİn-wİdth: 5in) { p {color: red; } }
+  </style>
+</head>
+
+<body>
+  <div><p lang="hi">स्टार</p></div>
+</body>
+</html>

layout/reftests/unicode/unicode-pseudo-selector.html

@@ -0,0 +1,14 @@
+<html>
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - pseudo element selectors</title>
+  <style>
+    p:fİrst-letter {color: red}
+  </style>
+</head>
+
+<body>
+  <div><p lang="hi">स्टार</p></div>
+  <input disabled>
+</body>
+</html>

layout/reftests/unicode/unicode-ref-print.html

@@ -0,0 +1,10 @@
+<html class="reftest-print">
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - media query - reference</title>
+</head>
+
+<body>
+  <div><p lang="hi">स्टार</p></div>
+</body>
+</html>

layout/reftests/unicode/unicode-ref.html

@@ -0,0 +1,11 @@
+<html>
+<head>
+  <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+  <title>Unicode tests - reference rending</title>
+</head>
+
+<body>
+  <div><p lang="hi">स्टार</p></div>
+  <input disabled>
+</body>
+</html>

layout/style/nsCSSParser.cpp

@@ -1645,7 +1645,7 @@ CSSParserImpl::ParseMediaQuery(PRUnichar aStopSymbol,
         return PR_FALSE;
       }
       // case insensitive from CSS - must be lower cased
-      ToLowerCase(mToken.mIdent);
+      nsContentUtils::ASCIIToLower(mToken.mIdent);
       mediaType = do_GetAtom(mToken.mIdent);
       if (gotNotOrOnly ||
           (mediaType != nsGkAtoms::_not && mediaType != nsGkAtoms::only))
@@ -1763,7 +1763,7 @@ CSSParserImpl::ParseMediaQueryExpression(nsMediaQuery* aQuery)
   }
 
   // case insensitive from CSS - must be lower cased
-  ToLowerCase(mToken.mIdent);
+  nsContentUtils::ASCIIToLower(mToken.mIdent);
   const PRUnichar *featureString;
   if (StringBeginsWith(mToken.mIdent, NS_LITERAL_STRING("min-"))) {
     expr->mRange = nsMediaExpression::eMin;
@@ -2922,7 +2922,7 @@ CSSParserImpl::ParseAttributeSelector(PRInt32&       aDataMask,
             short i = 0;
             const char* htmlAttr;
             while ((htmlAttr = caseInsensitiveHTMLAttribute[i++])) {
-              if (attr.EqualsIgnoreCase(htmlAttr)) {
+              if (attr.LowerCaseEqualsASCII(htmlAttr)) {
                 isCaseSensitive = PR_FALSE;
                 break;
               }
@@ -2996,7 +2996,7 @@ CSSParserImpl::ParsePseudoSelector(PRInt32&       aDataMask,
   nsAutoString buffer;
   buffer.Append(PRUnichar(':'));
   buffer.Append(mToken.mIdent);
-  ToLowerCase(buffer);
+  nsContentUtils::ASCIIToLower(buffer);
   nsCOMPtr<nsIAtom> pseudo = do_GetAtom(buffer);
   if (!pseudo) {
     mScanner.SetLowLevelError(NS_ERROR_OUT_OF_MEMORY);
@@ -3331,20 +3331,20 @@ CSSParserImpl::ParsePseudoClassWithNthPairArg(nsCSSSelector& aSelector,
   }
 
   if (eCSSToken_Ident == mToken.mType) {
-    if (mToken.mIdent.EqualsIgnoreCase("odd")) {
+    if (mToken.mIdent.LowerCaseEqualsLiteral("odd")) {
       numbers[0] = 2;
       numbers[1] = 1;
       lookForB = PR_FALSE;
     }
-    else if (mToken.mIdent.EqualsIgnoreCase("even")) {
+    else if (mToken.mIdent.LowerCaseEqualsLiteral("even")) {
       numbers[0] = 2;
       numbers[1] = 0;
       lookForB = PR_FALSE;
     }
-    else if (mToken.mIdent.EqualsIgnoreCase("n")) {
+    else if (mToken.mIdent.LowerCaseEqualsLiteral("n")) {
       numbers[0] = 1;
     }
-    else if (mToken.mIdent.EqualsIgnoreCase("-n")) {
+    else if (mToken.mIdent.LowerCaseEqualsLiteral("-n")) {
       numbers[0] = -1;
     }
     else {
@@ -3361,7 +3361,7 @@ CSSParserImpl::ParsePseudoClassWithNthPairArg(nsCSSSelector& aSelector,
     lookForB = PR_FALSE;
   }
   else if (eCSSToken_Dimension == mToken.mType) {
-    if (!mToken.mIntegerValid || !mToken.mIdent.EqualsIgnoreCase("n")) {
+    if (!mToken.mIntegerValid || !mToken.mIdent.LowerCaseEqualsLiteral("n")) {
       REPORT_UNEXPECTED_TOKEN(PEPseudoClassArgNotNth);
       return eSelectorParsingStatus_Error; // our caller calls SkipUntil(')')
     }

layout/style/nsCSSRuleProcessor.cpp

@@ -147,7 +147,7 @@ RuleHash_CIHashKey(PLDHashTable *table, const void *key)
 
   nsAutoString str;
   atom->ToString(str);
-  ToUpperCase(str);
+  nsContentUtils::ASCIIToLower(str);
   return HashString(str);
 }
 
@@ -1461,7 +1461,7 @@ static PRBool AttrMatchesValue(const nsAttrSelector* aAttrSelector,
     return PR_FALSE;
 
   const nsDefaultStringComparator defaultComparator;
-  const nsCaseInsensitiveStringComparator ciComparator;
+  const nsASCIICaseInsensitiveStringComparator ciComparator;
   const nsStringComparator& comparator =
       (aAttrSelector->mCaseSensitive || !isHTML)
                 ? static_cast<const nsStringComparator&>(defaultComparator)
@@ -1762,7 +1762,7 @@ static PRBool SelectorMatches(RuleProcessorData &data,
         if (lang && !lang->IsEmpty()) { // null check for out-of-memory
           if (!nsStyleUtil::DashMatchCompare(*lang,
                                              nsDependentString(pseudoClass->u.mString),
-                                             nsCaseInsensitiveStringComparator())) {
+                                             nsASCIICaseInsensitiveStringComparator())) {
             return PR_FALSE;
           }
           // This pseudo-class matched; move on to the next thing
@@ -1790,7 +1790,7 @@ static PRBool SelectorMatches(RuleProcessorData &data,
             if (nsStyleUtil::DashMatchCompare(Substring(language, begin,
                                                         end-begin),
                                               langString,
-                                              nsCaseInsensitiveStringComparator())) {
+                                              nsASCIICaseInsensitiveStringComparator())) {
               break;
             }
             begin = end + 1;

layout/style/nsCSSStyleRule.cpp

@@ -240,7 +240,7 @@ nsAttrSelector::nsAttrSelector(PRInt32 aNameSpace, const nsString& aAttr)
   MOZ_COUNT_CTOR(nsAttrSelector);
 
   nsAutoString lowercase;
-  ToLowerCase(aAttr, lowercase);
+  nsContentUtils::ASCIIToLower(aAttr, lowercase);
   
   mCasedAttr = do_GetAtom(aAttr);
   mLowercaseAttr = do_GetAtom(lowercase);
@@ -259,7 +259,7 @@ nsAttrSelector::nsAttrSelector(PRInt32 aNameSpace, const nsString& aAttr, PRUint
   MOZ_COUNT_CTOR(nsAttrSelector);
 
   nsAutoString lowercase;
-  ToLowerCase(aAttr, lowercase);
+  nsContentUtils::ASCIIToLower(aAttr, lowercase);
   
   mCasedAttr = do_GetAtom(aAttr);
   mLowercaseAttr = do_GetAtom(lowercase);
@@ -395,7 +395,7 @@ void nsCSSSelector::SetTag(const nsString& aTag)
   mCasedTag = do_GetAtom(aTag);
  
   nsAutoString lowercase;
-  ToLowerCase(aTag, lowercase);
+  nsContentUtils::ASCIIToLower(aTag, lowercase);
   mLowercaseTag = do_GetAtom(lowercase);
 }
 

layout/style/nsRuleNode.cpp

@@ -4569,7 +4569,7 @@ nsRuleNode::ComputeVisibilityData(void* aStartStruct,
       nsAutoString lang;
       displayData.mLang.GetStringValue(lang);
 
-      ToLowerCase(lang);
+      nsContentUtils::ASCIIToLower(lang);
       visibility->mLanguage = do_GetAtom(lang);
     }
   }