Changeset 61793 in webkit

Timestamp:

Jun 24, 2010 3:39:59 PM (14 years ago)

Author:

abarth@webkit.org

Message:

2010-06-24 Adam Barth <​abarth@webkit.org>

Reviewed by Eric Seidel.

Sketch out BeforeHTMLMode state
​https://bugs.webkit.org/show_bug.cgi?id=41133

In this patch, we switch the TreeBuidler over to using AtomicHTMLTokens
and sketch out another state. These states don't actually do anything
yet, I'm still just getting the structure of the spec into the code so
we can implement the details incrementally.

• html/HTMLTreeBuilder.cpp: (WebCore::HTMLTreeBuilder::constructTreeFromToken): (WebCore::HTMLTreeBuilder::processToken): (WebCore::HTMLTreeBuilder::insertDoctype): (WebCore::HTMLTreeBuilder::insertComment):
• html/HTMLTreeBuilder.h: (WebCore::HTMLTreeBuilder::parseError):

Location:

trunk/WebCore

Files:

• ChangeLog (modified) (1 diff)
• html/HTMLTreeBuilder.cpp (modified) (6 diffs)
• html/HTMLTreeBuilder.h (modified) (2 diffs)

trunk/WebCore/ChangeLog

@@ +1 @@
+2010-06-24  Adam Barth  <abarth@webkit.org>
+
+        Reviewed by Eric Seidel.
+
+        Sketch out BeforeHTMLMode state
+        https://bugs.webkit.org/show_bug.cgi?id=41133
+
+        In this patch, we switch the TreeBuidler over to using AtomicHTMLTokens
+        and sketch out another state.  These states don't actually do anything
+        yet, I'm still just getting the structure of the spec into the code so
+        we can implement the details incrementally.
+
+        * html/HTMLTreeBuilder.cpp:
+        (WebCore::HTMLTreeBuilder::constructTreeFromToken):
+        (WebCore::HTMLTreeBuilder::processToken):
+        (WebCore::HTMLTreeBuilder::insertDoctype):
+        (WebCore::HTMLTreeBuilder::insertComment):
+        * html/HTMLTreeBuilder.h:
+        (WebCore::HTMLTreeBuilder::parseError):
+
 2010-06-24  Adam Barth  <abarth@webkit.org>
 

trunk/WebCore/html/HTMLTreeBuilder.cpp

@@ -207 +207 @@
 }
 
-PassRefPtr<Node> HTMLTreeBuilder::constructTreeFromToken(HTMLToken& token)
+PassRefPtr<Node> HTMLTreeBuilder::constructTreeFromToken(HTMLToken& rawToken)
 {
     // Make MSVC ignore our unreachable code for now.
     if (true)
-        return passTokenToLegacyParser(token);
+        return passTokenToLegacyParser(rawToken);
+
+    AtomicHTMLToken token(rawToken);
 
     // HTML5 expects the tokenizer to call the parser every time a character is
@@ -218 +220 @@
     // the spec in only handling one character at a time.
     if (token.type() == HTMLToken::Character) {
-        HTMLToken::DataVector characters = token.characters();
-        HTMLToken::DataVector::const_iterator itr = characters.begin();
-        for (;itr; ++itr)
-            processToken(token, *itr);
+        StringImpl* characters = token.characters().impl();
+        // FIXME: Calling processToken for each character is probably slow.
+        for (unsigned i = 0; i < characters->length(); ++i)
+            processToken(token, (*characters)[i]);
         return 0; // FIXME: Should we be returning the Text node?
     }
@@ -227 +229 @@
 }
 
-PassRefPtr<Node> HTMLTreeBuilder::processToken(HTMLToken& token, UChar cc)
+PassRefPtr<Node> HTMLTreeBuilder::processToken(AtomicHTMLToken& token, UChar cc)
 {
 reprocessToken:
@@ -243 +245 @@
             if (cc == '\t' || cc == '\x0A' || cc == '\x0C' || cc == '\x0D' || cc == ' ')
                 return 0;
-            // Fall through
+            break;
         case HTMLToken::StartTag:
         case HTMLToken::EndTag:
         case HTMLToken::EndOfFile:
-            notImplemented();
-            setInsertionMode(BeforeHTMLMode);
-            goto reprocessToken;
+            break;
         }
-        ASSERT_NOT_REACHED();
-        break;
-    }
-    case BeforeHTMLMode:
+        notImplemented();
+        parseError(token);
+        setInsertionMode(BeforeHTMLMode);
+        goto reprocessToken;
+    }
+    case BeforeHTMLMode: {
+        switch (token.type()) {
+        case HTMLToken::Uninitialized:
+            ASSERT_NOT_REACHED();
+            break;
+        case HTMLToken::DOCTYPE:
+            parseError(token);
+            return 0;
+        case HTMLToken::Comment:
+            return insertComment(token);
+        case HTMLToken::Character:
+            if (cc == '\t' || cc == '\x0A' || cc == '\x0C' || cc == '\x0D' || cc == ' ')
+                return 0;
+            break;
+        case HTMLToken::StartTag:
+            if (token.name() == htmlTag) {
+                notImplemented();
+                setInsertionMode(BeforeHeadMode);
+                return 0;
+            }
+            break;
+        case HTMLToken::EndTag:
+            if (token.name() == headTag || token.name() == bodyTag || token.name() == htmlTag || token.name() == brTag)
+                break;
+            parseError(token);
+            return 0;
+        case HTMLToken::EndOfFile:
+            break;
+        }
+        notImplemented();
+        setInsertionMode(BeforeHeadMode);
+        goto reprocessToken;
+    }
     case BeforeHeadMode:
     case InHeadMode:
@@ -283 +317 @@
 }
 
-PassRefPtr<Node> HTMLTreeBuilder::insertDoctype(HTMLToken& token)
+PassRefPtr<Node> HTMLTreeBuilder::insertDoctype(AtomicHTMLToken& token)
 {
     ASSERT_UNUSED(token, token.type() == HTMLToken::DOCTYPE);
@@ -289 +323 @@
 }
 
-PassRefPtr<Node> HTMLTreeBuilder::insertComment(HTMLToken& token)
+PassRefPtr<Node> HTMLTreeBuilder::insertComment(AtomicHTMLToken& token)
 {
     ASSERT_UNUSED(token, token.type() == HTMLToken::Comment);

trunk/WebCore/html/HTMLTreeBuilder.h

@@ -37 +37 @@
 namespace WebCore {
 
+class AtomicHTMLToken;
 class Document;
 class DocumentFragment;
@@ -100 +101 @@
 
     PassRefPtr<Node> passTokenToLegacyParser(HTMLToken&);
-    PassRefPtr<Node> processToken(HTMLToken&, UChar cc = 0);
+    PassRefPtr<Node> processToken(AtomicHTMLToken&, UChar cc = 0);
 
-    PassRefPtr<Node> insertDoctype(HTMLToken&);
-    PassRefPtr<Node> insertComment(HTMLToken&);
+    PassRefPtr<Node> insertDoctype(AtomicHTMLToken&);
+    PassRefPtr<Node> insertComment(AtomicHTMLToken&);
+
+    // FIXME: Implement error reporting.
+    void parseError(AtomicHTMLToken&) { }
 
     void handleScriptStartTag();