Fixed partitioning bug causing an assertion exception and failing
[phpeclipse.git] / net.sourceforge.phpeclipse / src / net / sourceforge / phpeclipse / phpeditor / php / PHPPartitionScanner.java
index 81392ce..d1a64ad 100644 (file)
@@ -6,7 +6,7 @@
  * Created on 05.03.2003
  *
  * @author Stefan Langer (musk)
- * @version $Revision: 1.12 $
+ * @version $Revision: 1.16 $
  */
 package net.sourceforge.phpeclipse.phpeditor.php;
 
@@ -20,16 +20,19 @@ import org.eclipse.jface.text.rules.*;
  */
 public class PHPPartitionScanner implements IPartitionTokenScanner
 {
-    private static final boolean DEBUG = false;
+    private static final boolean DEBUG = true;
+    private boolean fInString = false;
+    private boolean fInDoubString = false;
     private IDocument fDocument = null;
     private int fOffset = -1;
     private String fContentType = IPHPPartitionScannerConstants.HTML;
-    private String fPrevContentType;
+    private String fPrevContentType = IPHPPartitionScannerConstants.HTML;
 
     private boolean partitionBorder = false;
     private int fTokenOffset;
     private int fEnd = -1;
     private int fLength;
+    private int fCurrentLength;
     private Map tokens = new HashMap();
 
     public PHPPartitionScanner()
@@ -53,7 +56,34 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
 
     private IToken getToken(String type)
     {
-        fLength = fOffset - fTokenOffset;
+        fLength = fCurrentLength;
+        if (DEBUG)
+        {
+
+            try
+            {
+                if (fLength <= 0)
+                {
+                    int line = fDocument.getLineOfOffset(fOffset);
+                    System.err.println(
+                        "Error at "
+                            + line
+                            + " offset:"
+                            + String.valueOf(
+                                fOffset - fDocument.getLineOffset(line)));
+                }
+            }
+            catch (BadLocationException e)
+            {
+                // TODO Auto-generated catch block
+                e.printStackTrace();
+            }
+        }
+        Assert.isTrue(fLength > 0, "Partition length <= 0!");
+        fCurrentLength = 0;
+        // String can never cross partition borders so reset string detection
+        fInString = false;
+        fInDoubString = false;
         IToken token = (IToken) this.tokens.get(type);
         Assert.isNotNull(token, "Token for type \"" + type + "\" not found!");
         if (DEBUG)
@@ -62,7 +92,7 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                 "Partition: fTokenOffset="
                     + fTokenOffset
                     + " fContentType="
-                    + fContentType
+                    + type
                     + " fLength="
                     + fLength);
         }
@@ -79,7 +109,6 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
         String contentType,
         int partitionOffset)
     {
-        this.setRange(document, offset, length);
         if (DEBUG)
         {
             System.out.println(
@@ -89,14 +118,35 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                     + partitionOffset);
         }
 
-        if (this.tokens.containsKey(contentType))
-            fContentType = contentType;
-        // TODO Calculate previouse contenttype
-        if (partitionOffset > -1)
+        try
         {
-            partitionBorder = false;
-            fTokenOffset = partitionOffset;
+            if (partitionOffset > -1)
+            {
+                partitionBorder = false;
+                // because of strings we have to parse the whole partition
+                this.setRange(
+                    document,
+                    partitionOffset,
+                    offset - partitionOffset + length);
+                // sometimes we get a wrong partition so we retrieve the partition
+                // directly from the document
+                fContentType = fDocument.getContentType(partitionOffset);
+                //TODO determine the previouse contenttypes as a stack
+                //if(partitionOffset > 1)
+                //     fPrevContentType = fDocument.getContentType(partitionOffset-1);
+            }
+            else
+                this.setRange(document, offset, length);
+
+        }
+        catch (BadLocationException e)
+        {
+            // should never happen
+            // TODO print stack trace to log
+            // fall back just scan the whole document again
+            this.setRange(document, 0, fDocument.getLength());
         }
+
     }
 
     /* (non-Javadoc)
@@ -143,18 +193,20 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
             switch (c)
             {
                 case '<' :
-                    if (fContentType
-                        != IPHPPartitionScannerConstants.PHP_MULTILINE_COMMENT
+                    if (!isInString(IPHPPartitionScannerConstants.PHP)
+                        && fContentType
+                            != IPHPPartitionScannerConstants.PHP_MULTILINE_COMMENT
                         && checkPattern(new char[] { '?', 'p', 'h', 'p' }, true))
                     {
-            if (fContentType != IPHPPartitionScannerConstants.PHP
-                            && fOffset - 5 > 0)
+                        if (fContentType != IPHPPartitionScannerConstants.PHP
+                            && fCurrentLength > 5)
                         {
-                            fOffset -= 5;
+                            unread(5);
                             IToken token = getToken(fContentType);
                             // save previouse contenttype
-              fPrevContentType = fContentType;
-                     
+                            //TODO build stack for previouse contenttype 
+                            fPrevContentType = fContentType;
+
                             fContentType = IPHPPartitionScannerConstants.PHP;
 
                             return token;
@@ -164,18 +216,42 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
 
                         // remember offset of this partition
                         fTokenOffset = fOffset - 5;
+                        fCurrentLength = 5;
                     }
-                    else if (checkPattern(new char[] { '!', '-', '-' }))
+                    else if (
+                        !isInString(IPHPPartitionScannerConstants.PHP)
+                            && fContentType
+                                != IPHPPartitionScannerConstants
+                                    .PHP_MULTILINE_COMMENT
+                            && checkPattern(new char[] { '?' }, false))
                     {
-                        // return previouse partition
+                        if (fContentType != IPHPPartitionScannerConstants.PHP
+                            && fCurrentLength > 2)
+                        {
+                            unread(2);
+                            IToken token = getToken(fContentType);
+                            // save previouse contenttype
+                            fPrevContentType = fContentType;
+                            fContentType = IPHPPartitionScannerConstants.PHP;
+                            return token;
+                        }
+                        else
+                            fContentType = IPHPPartitionScannerConstants.PHP;
+                        // remember offset of this partition
+                        fTokenOffset = fOffset - 2;
+                        fCurrentLength = 2;
+                    }
+                    else if (
+                        !isInString(IPHPPartitionScannerConstants.PHP)
+                            && checkPattern(new char[] { '!', '-', '-' }))
+                    { // return previouse partition
                         if (fContentType
                             != IPHPPartitionScannerConstants
                                 .HTML_MULTILINE_COMMENT
-                            && fOffset - 4 > 0)
+                            && fCurrentLength > 4)
                         {
-                            fOffset -= 4;
+                            unread(4);
                             IToken token = getToken(fContentType);
-
                             fContentType =
                                 IPHPPartitionScannerConstants
                                     .HTML_MULTILINE_COMMENT;
@@ -185,19 +261,22 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                             fContentType =
                                 IPHPPartitionScannerConstants
                                     .HTML_MULTILINE_COMMENT;
-
+                                    
                         fTokenOffset = fOffset - 4;
+                        fCurrentLength = 4;
                     }
                     break;
                 case '?' :
-                    if (fContentType == IPHPPartitionScannerConstants.PHP)
+                    if (!isInString(IPHPPartitionScannerConstants.PHP)
+                        && fContentType == IPHPPartitionScannerConstants.PHP)
                     {
                         if ((c = read()) == '>')
-                        {      // TODO Actually calculate the previouse contenttype from the document
-                          if(fPrevContentType != null)
-                            fContentType = fPrevContentType;
-                          else
-                              fContentType = IPHPPartitionScannerConstants.HTML;
+                        { 
+                            if (fPrevContentType != null)
+                                fContentType = fPrevContentType;
+                            else
+                                fContentType =
+                                    IPHPPartitionScannerConstants.HTML;
                             partitionBorder = true;
                             return getToken(IPHPPartitionScannerConstants.PHP);
                         }
@@ -206,8 +285,10 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                     }
                     break;
                 case '-' :
-                    if (fContentType
-                        == IPHPPartitionScannerConstants.HTML_MULTILINE_COMMENT
+                    if (!isInString(IPHPPartitionScannerConstants.PHP)
+                        && fContentType
+                            == IPHPPartitionScannerConstants
+                                .HTML_MULTILINE_COMMENT
                         && checkPattern(new char[] { '-', '>' }))
                     {
                         fContentType = IPHPPartitionScannerConstants.HTML;
@@ -218,18 +299,16 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                     }
                     break;
                 case '/' :
-                    if ((c = read()) == '*')
+                    if (!isInString(IPHPPartitionScannerConstants.PHP) && (c = read()) == '*')
                     { // MULTINE COMMENT JAVASCRIPT, CSS, PHP
                         if (fContentType == IPHPPartitionScannerConstants.PHP
-                            && fOffset - 2 > 0)
+                            && fCurrentLength > 2)
                         {
-                            fOffset -= 2;
+                            unread(2);
                             IToken token = getToken(fContentType);
-
                             fContentType =
                                 IPHPPartitionScannerConstants
                                     .PHP_MULTILINE_COMMENT;
-
                             return token;
                         }
                         else if (
@@ -239,14 +318,15 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                         {
 
                             fTokenOffset = fOffset - 2;
+                            fCurrentLength = 2;
                         }
 
                     }
-                    else if (c != ICharacterScanner.EOF)
+                    else if (!isInString(IPHPPartitionScannerConstants.PHP) && c != ICharacterScanner.EOF)
                         unread();
                     break;
                 case '*' :
-                    if ((c = read()) == '/')
+                    if (!isInString(IPHPPartitionScannerConstants.PHP) && (c = read()) == '/')
                     {
                         if (fContentType
                             == IPHPPartitionScannerConstants
@@ -254,7 +334,6 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                         {
                             fContentType = IPHPPartitionScannerConstants.PHP;
                             partitionBorder = true;
-
                             return getToken(
                                 IPHPPartitionScannerConstants
                                     .PHP_MULTILINE_COMMENT);
@@ -272,13 +351,20 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                         {
                         }
                     }
-                    else if (c != ICharacterScanner.EOF)
+                    else if (!isInString(IPHPPartitionScannerConstants.PHP) && c != ICharacterScanner.EOF)
                         unread();
                     break;
+                case '\'' :
+                    if (!fInDoubString)
+                        fInString = !fInString;
+                    break;
+                case '"' :
+                    // toggle String mode
+                    if (!fInString)
+                        fInDoubString = !fInDoubString;
+                    break;
             }
-        }
-
-        // end of file reached but we have to return the
+        } // end of file reached but we have to return the
         // last partition.
         return getToken(fContentType);
     }
@@ -296,8 +382,11 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
         fDocument = document;
         fOffset = offset;
         fTokenOffset = offset;
+        fCurrentLength = 0;
         fLength = 0;
         fEnd = fOffset + length;
+        fInString = false;
+        fInDoubString = false;
         //partitionBorder = false;
     }
 
@@ -307,6 +396,7 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
         {
             if (fOffset < fEnd)
             {
+               fCurrentLength++;
                 return fDocument.getChar(fOffset++);
             }
             return ICharacterScanner.EOF;
@@ -323,6 +413,13 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
     private void unread()
     {
         --fOffset;
+        --fCurrentLength;
+    }
+    
+    private void unread(int num)
+    {
+       fOffset -= num;
+       fCurrentLength -= num;
     }
 
     private boolean checkPattern(char[] pattern)
@@ -341,6 +438,7 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
     private boolean checkPattern(char[] pattern, boolean ignoreCase)
     {
         int prevOffset = fOffset;
+        int prevLength = fCurrentLength;
         for (int i = 0; i < pattern.length; i++)
         {
             int c = read();
@@ -349,6 +447,7 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
                 || !letterEquals(c, pattern[i], ignoreCase))
             {
                 fOffset = prevOffset;
+                fCurrentLength = prevLength;
                 return false;
             }
         }
@@ -373,6 +472,22 @@ public class PHPPartitionScanner implements IPartitionTokenScanner
 
         return false;
     }
+    
+    /**
+     * Checks wether the offset is in a <code>String</code> and the specified 
+     * contenttype is the current content type.
+     * Strings are delimited, mutual exclusive, by a " or by a '.
+     * 
+     * @param contentType The contenttype to check.
+     * @return <code>true</code> if the current offset is in a string else 
+     *                         returns false.
+     */
+    private  boolean isInString(String contentType)
+    {
+       if(fContentType == contentType)
+               return (fInString || fInDoubString);
+       else
+               return false;           
+    }
 
 }
-