refactory: added UI removed from core plugin.
[phpeclipse.git] / net.sourceforge.phpeclipse.ui / src / net / sourceforge / phpdt / internal / ui / text / JavaIndenter.java
diff --git a/net.sourceforge.phpeclipse.ui/src/net/sourceforge/phpdt/internal/ui/text/JavaIndenter.java b/net.sourceforge.phpeclipse.ui/src/net/sourceforge/phpdt/internal/ui/text/JavaIndenter.java
new file mode 100644 (file)
index 0000000..ef0c4ae
--- /dev/null
@@ -0,0 +1,1519 @@
+/*******************************************************************************
+ * Copyright (c) 2000, 2004 IBM Corporation and others.
+ * All rights reserved. This program and the accompanying materials 
+ * are made available under the terms of the Common Public License v1.0
+ * which accompanies this distribution, and is available at
+ * http://www.eclipse.org/legal/cpl-v10.html
+ * 
+ * Contributors:
+ *     IBM Corporation - initial API and implementation
+ *******************************************************************************/
+package net.sourceforge.phpdt.internal.ui.text;
+
+import net.sourceforge.phpdt.core.JavaCore;
+import net.sourceforge.phpdt.core.formatter.DefaultCodeFormatterConstants;
+import net.sourceforge.phpeclipse.ui.WebUI;
+//import net.sourceforge.phpeclipse.PHPeclipsePlugin;
+
+import org.eclipse.core.runtime.Plugin;
+import org.eclipse.jface.text.Assert;
+import org.eclipse.jface.text.BadLocationException;
+import org.eclipse.jface.text.IDocument;
+import org.eclipse.jface.text.IRegion;
+import org.eclipse.jface.text.ITypedRegion;
+import org.eclipse.jface.text.TextUtilities;
+import org.eclipse.ui.texteditor.AbstractDecoratedTextEditorPreferenceConstants;
+
+/**
+ * Uses the {@link net.sourceforge.phpdt.internal.ui.text.JavaHeuristicScanner}to
+ * get the indentation level for a certain position in a document.
+ * 
+ * <p>
+ * An instance holds some internal position in the document and is therefore not
+ * threadsafe.
+ * </p>
+ * 
+ * @since 3.0
+ */
+public class JavaIndenter {
+
+       /** The document being scanned. */
+       private IDocument fDocument;
+
+       /** The indentation accumulated by <code>findPreviousIndenationUnit</code>. */
+       private int fIndent;
+
+       /**
+        * The absolute (character-counted) indentation offset for special cases
+        * (method defs, array initializers)
+        */
+       private int fAlign;
+
+       /** The stateful scanposition for the indentation methods. */
+       private int fPosition;
+
+       /** The previous position. */
+       private int fPreviousPos;
+
+       /** The most recent token. */
+       private int fToken;
+
+       /** The line of <code>fPosition</code>. */
+       private int fLine;
+
+       /**
+        * The scanner we will use to scan the document. It has to be installed on
+        * the same document as the one we get.
+        */
+       private JavaHeuristicScanner fScanner;
+
+       /**
+        * Creates a new instance.
+        * 
+        * @param document
+        *            the document to scan
+        * @param scanner
+        *            the {@link JavaHeuristicScanner} to be used for scanning the
+        *            document. It must be installed on the same
+        *            <code>IDocument</code>.
+        */
+       public JavaIndenter(IDocument document, JavaHeuristicScanner scanner) {
+               Assert.isNotNull(document);
+               Assert.isNotNull(scanner);
+               fDocument = document;
+               fScanner = scanner;
+       }
+
+       /**
+        * Computes the indentation at the reference point of <code>position</code>.
+        * 
+        * @param offset
+        *            the offset in the document
+        * @return a String which reflects the indentation at the line in which the
+        *         reference position to <code>offset</code> resides, or
+        *         <code>null</code> if it cannot be determined
+        */
+       public StringBuffer getReferenceIndentation(int offset) {
+               return getReferenceIndentation(offset, false);
+       }
+
+       /**
+        * Computes the indentation at the reference point of <code>position</code>.
+        * 
+        * @param offset
+        *            the offset in the document
+        * @param assumeOpeningBrace
+        *            <code>true</code> if an opening brace should be assumed
+        * @return a String which reflects the indentation at the line in which the
+        *         reference position to <code>offset</code> resides, or
+        *         <code>null</code> if it cannot be determined
+        */
+       private StringBuffer getReferenceIndentation(int offset,
+                       boolean assumeOpeningBrace) {
+
+               int unit;
+               if (assumeOpeningBrace)
+                       unit = findReferencePosition(offset, Symbols.TokenLBRACE);
+               else
+                       unit = findReferencePosition(offset, peekChar(offset));
+
+               // if we were unable to find anything, return null
+               if (unit == JavaHeuristicScanner.NOT_FOUND)
+                       return null;
+
+               return getLeadingWhitespace(unit);
+
+       }
+
+       /**
+        * Computes the indentation at <code>offset</code>.
+        * 
+        * @param offset
+        *            the offset in the document
+        * @return a String which reflects the correct indentation for the line in
+        *         which offset resides, or <code>null</code> if it cannot be
+        *         determined
+        */
+       public StringBuffer computeIndentation(int offset) {
+               return computeIndentation(offset, false);
+       }
+
+       /**
+        * Computes the indentation at <code>offset</code>.
+        * 
+        * @param offset
+        *            the offset in the document
+        * @param assumeOpeningBrace
+        *            <code>true</code> if an opening brace should be assumed
+        * @return a String which reflects the correct indentation for the line in
+        *         which offset resides, or <code>null</code> if it cannot be
+        *         determined
+        */
+       public StringBuffer computeIndentation(int offset,
+                       boolean assumeOpeningBrace) {
+
+               StringBuffer indent = getReferenceIndentation(offset,
+                               assumeOpeningBrace);
+
+               // handle special alignment
+               if (fAlign != JavaHeuristicScanner.NOT_FOUND) {
+                       try {
+                               // a special case has been detected.
+                               IRegion line = fDocument.getLineInformationOfOffset(fAlign);
+                               int lineOffset = line.getOffset();
+                               return createIndent(lineOffset, fAlign);
+                       } catch (BadLocationException e) {
+                               return null;
+                       }
+               }
+
+               if (indent == null)
+                       return null;
+
+               // add additional indent
+               //indent.append(createIndent(fIndent));
+               indent.insert(0, createIndent(fIndent));
+               if (fIndent < 0)
+                       unindent(indent);
+
+               return indent;
+       }
+
+       /**
+        * Returns the indentation of the line at <code>offset</code> as a
+        * <code>StringBuffer</code>. If the offset is not valid, the empty
+        * string is returned.
+        * 
+        * @param offset
+        *            the offset in the document
+        * @return the indentation (leading whitespace) of the line in which
+        *         <code>offset</code> is located
+        */
+       private StringBuffer getLeadingWhitespace(int offset) {
+               StringBuffer indent = new StringBuffer();
+               try {
+                       IRegion line = fDocument.getLineInformationOfOffset(offset);
+                       int lineOffset = line.getOffset();
+                       int nonWS = fScanner.findNonWhitespaceForwardInAnyPartition(
+                                       lineOffset, lineOffset + line.getLength());
+                       indent.append(fDocument.get(lineOffset, nonWS - lineOffset));
+                       return indent;
+               } catch (BadLocationException e) {
+                       return indent;
+               }
+       }
+
+       /**
+        * Reduces indentation in <code>indent</code> by one indentation unit.
+        * 
+        * @param indent
+        *            the indentation to be modified
+        */
+       private void unindent(StringBuffer indent) {
+               CharSequence oneIndent = createIndent();
+               int i = indent.lastIndexOf(oneIndent.toString()); //$NON-NLS-1$
+               if (i != -1) {
+                       indent.delete(i, i + oneIndent.length());
+               }
+       }
+
+       /**
+        * Creates an indentation string of the length indent - start + 1,
+        * consisting of the content in <code>fDocument</code> in the range
+        * [start, indent), with every character replaced by a space except for
+        * tabs, which are kept as such.
+        * 
+        * <p>
+        * Every run of the number of spaces that make up a tab are replaced by a
+        * tab character.
+        * </p>
+        * 
+        * @return the indentation corresponding to the document content specified
+        *         by <code>start</code> and <code>indent</code>
+        */
+       private StringBuffer createIndent(int start, int indent) {
+               final int tabLen = prefTabLength();
+               StringBuffer ret = new StringBuffer();
+               try {
+                       int spaces = 0;
+                       while (start < indent) {
+
+                               char ch = fDocument.getChar(start);
+                               if (ch == '\t') {
+                                       ret.append('\t');
+                                       spaces = 0;
+                               } else if (tabLen == -1) {
+                                       ret.append(' ');
+                               } else {
+                                       spaces++;
+                                       if (spaces == tabLen) {
+                                               ret.append('\t');
+                                               spaces = 0;
+                                       }
+                               }
+
+                               start++;
+                       }
+                       // remainder
+                       if (spaces == tabLen)
+                               ret.append('\t');
+                       else
+                               while (spaces-- > 0)
+                                       ret.append(' ');
+
+               } catch (BadLocationException e) {
+               }
+
+               return ret;
+       }
+
+       /**
+        * Creates a string that represents the given number of indents (can be
+        * spaces or tabs..)
+        * 
+        * @param indent
+        *            the requested indentation level.
+        * 
+        * @return the indentation specified by <code>indent</code>
+        */
+       public StringBuffer createIndent(int indent) {
+               StringBuffer oneIndent = createIndent();
+
+               StringBuffer ret = new StringBuffer();
+               while (indent-- > 0)
+                       ret.append(oneIndent);
+
+               return ret;
+       }
+
+       /**
+        * Creates a string that represents one indent (can be spaces or tabs..)
+        * 
+        * @return one indentation
+        */
+       private StringBuffer createIndent() {
+               // get a sensible default when running without the infrastructure for
+               // testing
+               StringBuffer oneIndent = new StringBuffer();
+               // JavaCore plugin= JavaCore.getJavaCore();
+               WebUI plugin = WebUI.getDefault();
+               if (plugin == null) {
+                       oneIndent.append('\t');
+               } else {
+                       if (JavaCore.SPACE
+                                       .equals(JavaCore
+                                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_TAB_CHAR))) {
+                               int tabLen = Integer
+                                               .parseInt(JavaCore
+                                                               .getOption(DefaultCodeFormatterConstants.FORMATTER_TAB_SIZE));
+                               for (int i = 0; i < tabLen; i++)
+                                       oneIndent.append(' ');
+                       } else if (JavaCore.TAB
+                                       .equals(JavaCore
+                                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_TAB_CHAR)))
+                               oneIndent.append('\t');
+                       else
+                               oneIndent.append('\t'); // default
+               }
+               return oneIndent;
+       }
+
+       /**
+        * Returns the reference position regarding to indentation for
+        * <code>offset</code>, or <code>NOT_FOUND</code>. This method calls
+        * {@link #findReferencePosition(int, int) findReferencePosition(offset, nextChar)}
+        * where <code>nextChar</code> is the next character after
+        * <code>offset</code>.
+        * 
+        * @param offset
+        *            the offset for which the reference is computed
+        * @return the reference statement relative to which <code>offset</code>
+        *         should be indented, or {@link JavaHeuristicScanner#NOT_FOUND}
+        */
+       public int findReferencePosition(int offset) {
+               return findReferencePosition(offset, peekChar(offset));
+       }
+
+       /**
+        * Peeks the next char in the document that comes after <code>offset</code>
+        * on the same line as <code>offset</code>.
+        * 
+        * @param offset
+        *            the offset into document
+        * @return the token symbol of the next element, or TokenEOF if there is
+        *         none
+        */
+       private int peekChar(int offset) {
+               if (offset < fDocument.getLength()) {
+                       try {
+                               IRegion line = fDocument.getLineInformationOfOffset(offset);
+                               int lineOffset = line.getOffset();
+                               int next = fScanner.nextToken(offset, lineOffset
+                                               + line.getLength());
+                               return next;
+                       } catch (BadLocationException e) {
+                       }
+               }
+               return Symbols.TokenEOF;
+       }
+
+       /**
+        * Returns the reference position regarding to indentation for
+        * <code>position</code>, or <code>NOT_FOUND</code>.
+        * 
+        * <p>
+        * If <code>peekNextChar</code> is <code>true</code>, the next token
+        * after <code>offset</code> is read and taken into account when computing
+        * the indentation. Currently, if the next token is the first token on the
+        * line (i.e. only preceded by whitespace), the following tokens are
+        * specially handled:
+        * <ul>
+        * <li><code>switch</code> labels are indented relative to the switch
+        * block</li>
+        * <li>opening curly braces are aligned correctly with the introducing code</li>
+        * <li>closing curly braces are aligned properly with the introducing code
+        * of the matching opening brace</li>
+        * <li>closing parenthesis' are aligned with their opening peer</li>
+        * <li>the <code>else</code> keyword is aligned with its <code>if</code>,
+        * anything else is aligned normally (i.e. with the base of any introducing
+        * statements).</li>
+        * <li>if there is no token on the same line after <code>offset</code>,
+        * the indentation is the same as for an <code>else</code> keyword</li>
+        * </ul>
+        * 
+        * @param offset
+        *            the offset for which the reference is computed
+        * @param nextToken
+        *            the next token to assume in the document
+        * @return the reference statement relative to which <code>offset</code>
+        *         should be indented, or {@link JavaHeuristicScanner#NOT_FOUND}
+        */
+       public int findReferencePosition(int offset, int nextToken) {
+               boolean danglingElse = false;
+               boolean unindent = false;
+               boolean indent = false;
+               boolean matchBrace = false;
+               boolean matchParen = false;
+               boolean matchCase = false;
+
+               // account for unindenation characters already typed in, but after
+               // position
+               // if they are on a line by themselves, the indentation gets adjusted
+               // accordingly
+               //
+               // also account for a dangling else
+               if (offset < fDocument.getLength()) {
+                       try {
+                               IRegion line = fDocument.getLineInformationOfOffset(offset);
+                               int lineOffset = line.getOffset();
+                               int prevPos = Math.max(offset - 1, 0);
+                               boolean isFirstTokenOnLine = fDocument.get(lineOffset,
+                                               prevPos + 1 - lineOffset).trim().length() == 0;
+                               int prevToken = fScanner.previousToken(prevPos,
+                                               JavaHeuristicScanner.UNBOUND);
+                               if (prevToken == Symbols.TokenEOF && nextToken == Symbols.TokenEOF) {
+                                       ITypedRegion partition = TextUtilities.getPartition(fDocument, IPHPPartitions.PHP_PARTITIONING, offset, true);
+                                       if (partition.getType().equals(IPHPPartitions.PHP_SINGLELINE_COMMENT)) {
+                                               fAlign = fScanner.getPosition();
+                                       } else {
+                                               fAlign = JavaHeuristicScanner.NOT_FOUND;
+                                       }
+                                       return JavaHeuristicScanner.NOT_FOUND;
+                               }
+                               boolean bracelessBlockStart = fScanner.isBracelessBlockStart(
+                                               prevPos, JavaHeuristicScanner.UNBOUND);
+
+                               switch (nextToken) {
+                               case Symbols.TokenEOF:
+                               case Symbols.TokenELSE:
+                                       danglingElse = true;
+                                       break;
+                               case Symbols.TokenCASE:
+                               case Symbols.TokenDEFAULT:
+                                       if (isFirstTokenOnLine)
+                                               matchCase = true;
+                                       break;
+                               case Symbols.TokenLBRACE: // for opening-brace-on-new-line
+                                                                                       // style
+                               // if (bracelessBlockStart && !prefIndentBracesForBlocks())
+                               // unindent= true;
+                               // else if ((prevToken == Symbols.TokenCOLON || prevToken ==
+                               // Symbols.TokenEQUAL || prevToken == Symbols.TokenRBRACKET) &&
+                               // !prefIndentBracesForArrays())
+                               // unindent= true;
+                               // else if (!bracelessBlockStart &&
+                               // prefIndentBracesForMethods())
+                               // indent= true;
+                               // break;
+                                       if (bracelessBlockStart)
+                                               unindent = true;
+                                       else if ((prevToken == Symbols.TokenCOLON
+                                                       || prevToken == Symbols.TokenEQUAL || prevToken == Symbols.TokenRBRACKET))
+                                               unindent = true;
+                                       else if (!bracelessBlockStart)
+                                               indent = true;
+                                       break;
+                               case Symbols.TokenRBRACE: // closing braces get unindented
+                                       if (isFirstTokenOnLine)
+                                               matchBrace = true;
+                                       break;
+                               case Symbols.TokenRPAREN:
+                                       if (isFirstTokenOnLine)
+                                               matchParen = true;
+                                       break;
+                               }
+                       } catch (BadLocationException e) {
+                       }
+               } else {
+                       // assume an else could come if we are at the end of file
+                       danglingElse = true;
+               }
+
+               int ref = findReferencePosition(offset, danglingElse, matchBrace,
+                               matchParen, matchCase);
+               if (unindent)
+                       fIndent--;
+               if (indent)
+                       fIndent++;
+               return ref;
+       }
+
+       /**
+        * Returns the reference position regarding to indentation for
+        * <code>position</code>, or <code>NOT_FOUND</code>.<code>fIndent</code>
+        * will contain the relative indentation (in indentation units, not
+        * characters) after the call. If there is a special alignment (e.g. for a
+        * method declaration where parameters should be aligned),
+        * <code>fAlign</code> will contain the absolute position of the alignment
+        * reference in <code>fDocument</code>, otherwise <code>fAlign</code>
+        * is set to <code>JavaHeuristicScanner.NOT_FOUND</code>.
+        * 
+        * @param offset
+        *            the offset for which the reference is computed
+        * @param danglingElse
+        *            whether a dangling else should be assumed at
+        *            <code>position</code>
+        * @param matchBrace
+        *            whether the position of the matching brace should be returned
+        *            instead of doing code analysis
+        * @param matchParen
+        *            whether the position of the matching parenthesis should be
+        *            returned instead of doing code analysis
+        * @param matchCase
+        *            whether the position of a switch statement reference should be
+        *            returned (either an earlier case statement or the switch block
+        *            brace)
+        * @return the reference statement relative to which <code>position</code>
+        *         should be indented, or {@link JavaHeuristicScanner#NOT_FOUND}
+        */
+       public int findReferencePosition(int offset, boolean danglingElse,
+                       boolean matchBrace, boolean matchParen, boolean matchCase) {
+               fIndent = 0; // the indentation modification
+               fAlign = JavaHeuristicScanner.NOT_FOUND;
+               fPosition = offset;
+
+               // forward cases
+               // an unindentation happens sometimes if the next token is special,
+               // namely on braces, parens and case labels
+               // align braces, but handle the case where we align with the method
+               // declaration start instead of
+               // the opening brace.
+               if (matchBrace) {
+                       if (skipScope(Symbols.TokenLBRACE, Symbols.TokenRBRACE)) {
+                               try {
+                                       // align with the opening brace that is on a line by its own
+                                       int lineOffset = fDocument.getLineOffset(fLine);
+                                       if (lineOffset <= fPosition
+                                                       && fDocument
+                                                                       .get(lineOffset, fPosition - lineOffset)
+                                                                       .trim().length() == 0)
+                                               return fPosition;
+                               } catch (BadLocationException e) {
+                                       // concurrent modification - walk default path
+                               }
+                               // if the opening brace is not on the start of the line, skip to
+                               // the start
+                               int pos = skipToStatementStart(true, true);
+                               fIndent = 0; // indent is aligned with reference position
+                               return pos;
+                       } else {
+                               // if we can't find the matching brace, the heuristic is to
+                               // unindent
+                               // by one against the normal position
+                               int pos = findReferencePosition(offset, danglingElse, false,
+                                               matchParen, matchCase);
+                               fIndent--;
+                               return pos;
+                       }
+               }
+
+               // align parenthesis'
+               if (matchParen) {
+                       if (skipScope(Symbols.TokenLPAREN, Symbols.TokenRPAREN))
+                               return fPosition;
+                       else {
+                               // if we can't find the matching paren, the heuristic is to
+                               // unindent
+                               // by one against the normal position
+                               int pos = findReferencePosition(offset, danglingElse,
+                                               matchBrace, false, matchCase);
+                               fIndent--;
+                               return pos;
+                       }
+               }
+
+               // the only reliable way to get case labels aligned (due to many
+               // different styles of using braces in a block)
+               // is to go for another case statement, or the scope opening brace
+               if (matchCase) {
+                       return matchCaseAlignment();
+               }
+
+               nextToken();
+               switch (fToken) {
+               case Symbols.TokenRBRACE:
+                       // skip the block and fall through
+                       // if we can't complete the scope, reset the scan position
+                       int pos = fPosition;
+                       if (!skipScope())
+                               fPosition = pos;
+               case Symbols.TokenSEMICOLON:
+                       // this is the 90% case: after a statement block
+                       // the end of the previous statement / block previous.end
+                       // search to the end of the statement / block before the previous;
+                       // the token just after that is previous.start
+                       return skipToStatementStart(danglingElse, false);
+
+                       // scope introduction: special treat who special is
+               case Symbols.TokenLPAREN:
+               case Symbols.TokenLBRACE:
+               case Symbols.TokenLBRACKET:
+                       return handleScopeIntroduction(offset + 1);
+
+               case Symbols.TokenEOF:
+                       // trap when hitting start of document
+                       return 0;
+
+               case Symbols.TokenEQUAL:
+                       // indent assignments
+                       fIndent = prefAssignmentIndent();
+                       return fPosition;
+
+               case Symbols.TokenCOLON:
+                       // TODO handle ternary deep indentation
+                       fIndent = prefCaseBlockIndent();
+                       return fPosition;
+
+               case Symbols.TokenQUESTIONMARK:
+                       if (prefTernaryDeepAlign()) {
+                               setFirstElementAlignment(fPosition, offset + 1);
+                               return fPosition;
+                       } else {
+                               fIndent = prefTernaryIndent();
+                               return fPosition;
+                       }
+
+                       // indentation for blockless introducers:
+               case Symbols.TokenDO:
+               case Symbols.TokenWHILE:
+               case Symbols.TokenELSE:
+                       fIndent = prefSimpleIndent();
+                       return fPosition;
+               case Symbols.TokenRPAREN:
+                       int line = fLine;
+                       if (skipScope(Symbols.TokenLPAREN, Symbols.TokenRPAREN)) {
+                               int scope = fPosition;
+                               nextToken();
+                               if (fToken == Symbols.TokenIF || fToken == Symbols.TokenWHILE
+                                               || fToken == Symbols.TokenFOR) {
+                                       fIndent = prefSimpleIndent();
+                                       return fPosition;
+                               }
+                               fPosition = scope;
+                               if (looksLikeMethodDecl()) {
+                                       return skipToStatementStart(danglingElse, false);
+                               }
+                       }
+                       // restore
+                       fPosition = offset;
+                       fLine = line;
+                       // else: fall through to default
+
+               case Symbols.TokenCOMMA:
+                       // inside a list of some type
+                       // easy if there is already a list item before with its own
+                       // indentation - we just align
+                       // if not: take the start of the list ( LPAREN, LBRACE, LBRACKET )
+                       // and either align or
+                       // indent by list-indent
+               default:
+                       // inside whatever we don't know about: similar to the list case:
+                       // if we are inside a continued expression, then either align with a
+                       // previous line that has indentation
+                       // or indent from the expression start line (either a scope
+                       // introducer or the start of the expr).
+                       return skipToPreviousListItemOrListStart();
+
+               }
+       }
+
+       /**
+        * Skips to the start of a statement that ends at the current position.
+        * 
+        * @param danglingElse
+        *            whether to indent aligned with the last <code>if</code>
+        * @param isInBlock
+        *            whether the current position is inside a block, which limits
+        *            the search scope to the next scope introducer
+        * @return the reference offset of the start of the statement
+        */
+       private int skipToStatementStart(boolean danglingElse, boolean isInBlock) {
+               while (true) {
+                       nextToken();
+
+                       if (isInBlock) {
+                               switch (fToken) {
+                               // exit on all block introducers
+                               case Symbols.TokenIF:
+                               case Symbols.TokenELSE:
+                               case Symbols.TokenSYNCHRONIZED:
+                               case Symbols.TokenCOLON:
+                               case Symbols.TokenSTATIC:
+                               case Symbols.TokenCATCH:
+                               case Symbols.TokenDO:
+                               case Symbols.TokenWHILE:
+                               case Symbols.TokenFINALLY:
+                               case Symbols.TokenFOR:
+                               case Symbols.TokenTRY:
+                                       return fPosition;
+
+                               case Symbols.TokenSWITCH:
+                                       fIndent = prefCaseIndent();
+                                       return fPosition;
+                               }
+                       }
+
+                       switch (fToken) {
+                       // scope introduction through: LPAREN, LBRACE, LBRACKET
+                       // search stop on SEMICOLON, RBRACE, COLON, EOF
+                       // -> the next token is the start of the statement (i.e. previousPos
+                       // when backward scanning)
+                       case Symbols.TokenLPAREN:
+                       case Symbols.TokenLBRACE:
+                       case Symbols.TokenLBRACKET:
+                       case Symbols.TokenSEMICOLON:
+                       case Symbols.TokenEOF:
+                               return fPreviousPos;
+
+                       case Symbols.TokenCOLON:
+                               int pos = fPreviousPos;
+                               if (!isConditional())
+                                       return pos;
+                               break;
+
+                       case Symbols.TokenRBRACE:
+                               // RBRACE is a little tricky: it can be the end of an array
+                               // definition, but
+                               // usually it is the end of a previous block
+                               pos = fPreviousPos; // store state
+                               if (skipScope() && looksLikeArrayInitializerIntro())
+                                       continue; // it's an array
+                               else
+                                       return pos; // it's not - do as with all the above
+
+                               // scopes: skip them
+                       case Symbols.TokenRPAREN:
+                       case Symbols.TokenRBRACKET:
+                               pos = fPreviousPos;
+                               if (skipScope())
+                                       break;
+                               else
+                                       return pos;
+
+                               // IF / ELSE: align the position after the conditional block
+                               // with the if
+                               // so we are ready for an else, except if danglingElse is false
+                               // in order for this to work, we must skip an else to its if
+                       case Symbols.TokenIF:
+                               if (danglingElse)
+                                       return fPosition;
+                               else
+                                       break;
+                       case Symbols.TokenELSE:
+                               // skip behind the next if, as we have that one covered
+                               pos = fPosition;
+                               if (skipNextIF())
+                                       break;
+                               else
+                                       return pos;
+
+                       case Symbols.TokenDO:
+                               // align the WHILE position with its do
+                               return fPosition;
+
+                       case Symbols.TokenWHILE:
+                               // this one is tricky: while can be the start of a while loop
+                               // or the end of a do - while
+                               pos = fPosition;
+                               if (hasMatchingDo()) {
+                                       // continue searching from the DO on
+                                       break;
+                               } else {
+                                       // continue searching from the WHILE on
+                                       fPosition = pos;
+                                       break;
+                               }
+                       default:
+                               // keep searching
+
+                       }
+
+               }
+       }
+
+       /**
+        * Returns true if the colon at the current position is part of a
+        * conditional (ternary) expression, false otherwise.
+        * 
+        * @return true if the colon at the current position is part of a
+        *         conditional
+        */
+       private boolean isConditional() {
+               while (true) {
+                       nextToken();
+                       switch (fToken) {
+
+                       // search for case, otherwise return true
+                       case Symbols.TokenIDENT:
+                               continue;
+                       case Symbols.TokenCASE:
+                               return false;
+
+                       default:
+                               return true;
+                       }
+               }
+       }
+
+       /**
+        * Returns as a reference any previous <code>switch</code> labels (<code>case</code>
+        * or <code>default</code>) or the offset of the brace that scopes the
+        * switch statement. Sets <code>fIndent</code> to
+        * <code>prefCaseIndent</code> upon a match.
+        * 
+        * @return the reference offset for a <code>switch</code> label
+        */
+       private int matchCaseAlignment() {
+               while (true) {
+                       nextToken();
+                       switch (fToken) {
+                       // invalid cases: another case label or an LBRACE must come before a
+                       // case
+                       // -> bail out with the current position
+                       case Symbols.TokenLPAREN:
+                       case Symbols.TokenLBRACKET:
+                       case Symbols.TokenEOF:
+                               return fPosition;
+                       case Symbols.TokenLBRACE:
+                               // opening brace of switch statement
+                               fIndent = 1; //prefCaseIndent() is for Java
+                               return fPosition;
+                       case Symbols.TokenCASE:
+                       case Symbols.TokenDEFAULT:
+                               // align with previous label
+                               fIndent = 0;
+                               return fPosition;
+                       // scopes: skip them
+                       case Symbols.TokenRPAREN:
+                       case Symbols.TokenRBRACKET:
+                       case Symbols.TokenRBRACE:
+                               skipScope();
+                               break;
+                       default:
+                               // keep searching
+                               continue;
+                       }
+               }
+       }
+
+        /**
+        * Returns the reference position for a list element. The algorithm tries to
+        * match any previous indentation on the same list. If there is none, the
+        * reference position returned is determined depending on the type of list:
+        * The indentation will either match the list scope introducer (e.g. for
+        * method declarations), so called deep indents, or simply increase the
+        * indentation by a number of standard indents. See also
+        * {@link #handleScopeIntroduction(int)}.
+        * 
+        * @return the reference position for a list item: either a previous list
+        *         item that has its own indentation, or the list introduction
+        *         start.
+        */
+       private int skipToPreviousListItemOrListStart() {
+               int startLine = fLine;
+               int startPosition = fPosition;
+               while (true) {
+                       nextToken();
+
+                       // if any line item comes with its own indentation, adapt to it
+                       if (fLine < startLine) {
+                               try {
+                                       int lineOffset = fDocument.getLineOffset(startLine);
+                                       int bound = Math.min(fDocument.getLength(),
+                                                       startPosition + 1);
+                                       fAlign = fScanner.findNonWhitespaceForwardInAnyPartition(
+                                                       lineOffset, bound);
+                               } catch (BadLocationException e) {
+                                       // ignore and return just the position
+                               }
+                               return startPosition;
+                       }
+
+                       switch (fToken) {
+                       // scopes: skip them
+                       case Symbols.TokenRPAREN:
+                       case Symbols.TokenRBRACKET:
+                       case Symbols.TokenRBRACE:
+                               skipScope();
+                               break;
+
+                       // scope introduction: special treat who special is
+                       case Symbols.TokenLPAREN:
+                       case Symbols.TokenLBRACE:
+                       case Symbols.TokenLBRACKET:
+                               return handleScopeIntroduction(startPosition + 1);
+
+                       case Symbols.TokenSEMICOLON:
+                               return fPosition;
+                       case Symbols.TokenQUESTIONMARK:
+                               if (prefTernaryDeepAlign()) {
+                                       setFirstElementAlignment(fPosition - 1, fPosition + 1);
+                               } else {
+                                       fIndent = prefTernaryIndent();
+                               }
+                               return fPosition;
+                       case Symbols.TokenEOF:
+                               return 0;
+
+                       case Symbols.TokenEQUAL:
+                               // indent assignments
+                               fIndent= prefAssignmentIndent();
+                               return fPosition;
+                       }
+               }
+       }
+
+       /**
+        * Skips a scope and positions the cursor (<code>fPosition</code>) on
+        * the token that opens the scope. Returns <code>true</code> if a matching
+        * peer could be found, <code>false</code> otherwise. The current token
+        * when calling must be one out of <code>Symbols.TokenRPAREN</code>,
+        * <code>Symbols.TokenRBRACE</code>, and
+        * <code>Symbols.TokenRBRACKET</code>.
+        * 
+        * @return <code>true</code> if a matching peer was found,
+        *         <code>false</code> otherwise
+        */
+       private boolean skipScope() {
+               switch (fToken) {
+               case Symbols.TokenRPAREN:
+                       return skipScope(Symbols.TokenLPAREN, Symbols.TokenRPAREN);
+               case Symbols.TokenRBRACKET:
+                       return skipScope(Symbols.TokenLBRACKET, Symbols.TokenRBRACKET);
+               case Symbols.TokenRBRACE:
+                       return skipScope(Symbols.TokenLBRACE, Symbols.TokenRBRACE);
+               default:
+                       Assert.isTrue(false);
+                       return false;
+               }
+       }
+
+       /**
+        * Handles the introduction of a new scope. The current token must be one
+        * out of <code>Symbols.TokenLPAREN</code>,
+        * <code>Symbols.TokenLBRACE</code>, and
+        * <code>Symbols.TokenLBRACKET</code>. Returns as the reference position
+        * either the token introducing the scope or - if available - the first java
+        * token after that.
+        * 
+        * <p>
+        * Depending on the type of scope introduction, the indentation will align
+        * (deep indenting) with the reference position (<code>fAlign</code> will
+        * be set to the reference position) or <code>fIndent</code> will be set
+        * to the number of indentation units.
+        * </p>
+        * 
+        * @param bound
+        *            the bound for the search for the first token after the scope
+        *            introduction.
+        * @return
+        */
+       private int handleScopeIntroduction(int bound) {
+               switch (fToken) {
+               // scope introduction: special treat who special is
+               case Symbols.TokenLPAREN:
+                       int pos = fPosition; // store
+
+                       // special: method declaration deep indentation
+                       if (looksLikeMethodDecl()) {
+                               if (prefMethodDeclDeepIndent())
+                                       return setFirstElementAlignment(pos, bound);
+                               else {
+                                       fIndent = prefMethodDeclIndent();
+                                       return pos;
+                               }
+                       } else {
+                               fPosition = pos;
+                               if (looksLikeMethodCall()) {
+                                       if (prefMethodCallDeepIndent())
+                                               return setFirstElementAlignment(pos, bound);
+                                       else {
+                                               fIndent = prefMethodCallIndent();
+                                               return pos;
+                                       }
+                               } else if (prefParenthesisDeepIndent())
+                                       return setFirstElementAlignment(pos, bound);
+                       }
+
+                       // normal: return the parenthesis as reference
+                       fIndent = prefParenthesisIndent();
+                       return pos;
+
+               case Symbols.TokenLBRACE:
+                       pos = fPosition; // store
+
+                       // special: array initializer
+                       if (looksLikeArrayInitializerIntro())
+                               if (prefArrayDeepIndent())
+                                       return setFirstElementAlignment(pos, bound);
+                               else
+                                       fIndent = prefArrayIndent();
+                       else
+                               fIndent = prefBlockIndent();
+
+                       // normal: skip to the statement start before the scope introducer
+                       // opening braces are often on differently ending indents than e.g.
+                       // a method definition
+                       fPosition = pos; // restore
+                       return skipToStatementStart(true, true); // set to true to match
+                                                                                                               // the first if
+
+               case Symbols.TokenLBRACKET:
+                       pos = fPosition; // store
+
+                       // special: method declaration deep indentation
+                       if (prefArrayDimensionsDeepIndent()) {
+                               return setFirstElementAlignment(pos, bound);
+                       }
+
+                       // normal: return the bracket as reference
+                       fIndent = prefBracketIndent();
+                       return pos; // restore
+
+               default:
+                       Assert.isTrue(false);
+                       return -1; // dummy
+               }
+       }
+
+       /**
+        * Sets the deep indent offset (<code>fAlign</code>) to either the
+        * offset right after <code>scopeIntroducerOffset</code> or - if available -
+        * the first Java token after <code>scopeIntroducerOffset</code>, but
+        * before <code>bound</code>.
+        * 
+        * @param scopeIntroducerOffset
+        *            the offset of the scope introducer
+        * @param bound
+        *            the bound for the search for another element
+        * @return the reference position
+        */
+       private int setFirstElementAlignment(int scopeIntroducerOffset, int bound) {
+               int firstPossible = scopeIntroducerOffset + 1; // align with the first
+                                                                                                               // position after the
+                                                                                                               // scope intro
+               fAlign = fScanner.findNonWhitespaceForwardInAnyPartition(firstPossible,
+                               bound);
+               if (fAlign == JavaHeuristicScanner.NOT_FOUND)
+                       fAlign = firstPossible;
+               return fAlign;
+       }
+
+       /**
+        * Returns <code>true</code> if the next token received after calling
+        * <code>nextToken</code> is either an equal sign or an array designator
+        * ('[]').
+        * 
+        * @return <code>true</code> if the next elements look like the start of
+        *         an array definition
+        */
+       private boolean looksLikeArrayInitializerIntro() {
+               nextToken();
+               if (fToken == Symbols.TokenEQUAL || skipBrackets()) {
+                       return true;
+               }
+               return false;
+       }
+
+       /**
+        * Skips over the next <code>if</code> keyword. The current token when
+        * calling this method must be an <code>else</code> keyword. Returns
+        * <code>true</code> if a matching <code>if</code> could be found,
+        * <code>false</code> otherwise. The cursor (<code>fPosition</code>)
+        * is set to the offset of the <code>if</code> token.
+        * 
+        * @return <code>true</code> if a matching <code>if</code> token was
+        *         found, <code>false</code> otherwise
+        */
+       private boolean skipNextIF() {
+               Assert.isTrue(fToken == Symbols.TokenELSE);
+
+               while (true) {
+                       nextToken();
+                       switch (fToken) {
+                       // scopes: skip them
+                       case Symbols.TokenRPAREN:
+                       case Symbols.TokenRBRACKET:
+                       case Symbols.TokenRBRACE:
+                               skipScope();
+                               break;
+
+                       case Symbols.TokenIF:
+                               // found it, return
+                               return true;
+                       case Symbols.TokenELSE:
+                               // recursively skip else-if blocks
+                               skipNextIF();
+                               break;
+
+                       // shortcut scope starts
+                       case Symbols.TokenLPAREN:
+                       case Symbols.TokenLBRACE:
+                       case Symbols.TokenLBRACKET:
+                       case Symbols.TokenEOF:
+                               return false;
+                       }
+               }
+       }
+
+       /**
+        * while(condition); is ambiguous when parsed backwardly, as it is a valid
+        * statement by its own, so we have to check whether there is a matching do.
+        * A <code>do</code> can either be separated from the while by a block, or
+        * by a single statement, which limits our search distance.
+        * 
+        * @return <code>true</code> if the <code>while</code> currently in
+        *         <code>fToken</code> has a matching <code>do</code>.
+        */
+       private boolean hasMatchingDo() {
+               Assert.isTrue(fToken == Symbols.TokenWHILE);
+               nextToken();
+               switch (fToken) {
+               case Symbols.TokenRBRACE:
+                       skipScope(); // and fall thru
+               case Symbols.TokenSEMICOLON:
+                       skipToStatementStart(false, false);
+                       return fToken == Symbols.TokenDO;
+               }
+               return false;
+       }
+
+       /**
+        * Skips brackets if the current token is a RBRACKET. There can be nothing
+        * but whitespace in between, this is only to be used for <code>[]</code>
+        * elements.
+        * 
+        * @return <code>true</code> if a <code>[]</code> could be scanned, the
+        *         current token is left at the LBRACKET.
+        */
+       private boolean skipBrackets() {
+               if (fToken == Symbols.TokenRBRACKET) {
+                       nextToken();
+                       if (fToken == Symbols.TokenLBRACKET) {
+                               return true;
+                       }
+               }
+               return false;
+       }
+
+       /**
+        * Reads the next token in backward direction from the heuristic scanner and
+        * sets the fields <code>fToken, fPreviousPosition</code> and
+        * <code>fPosition</code> accordingly.
+        */
+       private void nextToken() {
+               nextToken(fPosition);
+       }
+
+       /**
+        * Reads the next token in backward direction of <code>start</code> from
+        * the heuristic scanner and sets the fields
+        * <code>fToken, fPreviousPosition</code> and <code>fPosition</code>
+        * accordingly.
+        */
+       private void nextToken(int start) {
+               fToken = fScanner
+                               .previousToken(start - 1, JavaHeuristicScanner.UNBOUND);
+               fPreviousPos = start;
+               fPosition = fScanner.getPosition() + 1;
+               try {
+                       fLine = fDocument.getLineOfOffset(fPosition);
+               } catch (BadLocationException e) {
+                       fLine = -1;
+               }
+       }
+
+       /**
+        * Returns <code>true</code> if the current tokens look like a method
+        * declaration header (i.e. only the return type and method name). The
+        * heuristic calls <code>nextToken</code> and expects an identifier
+        * (method name) and a type declaration (an identifier with optional
+        * brackets) which also covers the visibility modifier of constructors; it
+        * does not recognize package visible constructors.
+        * 
+        * @return <code>true</code> if the current position looks like a method
+        *         declaration header.
+        */
+       private boolean looksLikeMethodDecl() {
+               /*
+                * TODO This heuristic does not recognize package private constructors
+                * since those do have neither type nor visibility keywords. One option
+                * would be to go over the parameter list, but that might be empty as
+                * well - hard to do without an AST...
+                */
+
+               nextToken();
+               if (fToken == Symbols.TokenIDENT) { // method name
+                       do
+                               nextToken();
+                       while (skipBrackets()); // optional brackets for array valued return
+                                                                       // types
+                       return fToken == Symbols.TokenIDENT; // type name
+
+               }
+               return false;
+       }
+
+       /**
+        * Returns <code>true</code> if the current tokens look like a method call
+        * header (i.e. an identifier as opposed to a keyword taking parenthesized
+        * parameters such as <code>if</code>).
+        * <p>
+        * The heuristic calls <code>nextToken</code> and expects an identifier
+        * (method name).
+        * 
+        * @return <code>true</code> if the current position looks like a method
+        *         call header.
+        */
+       private boolean looksLikeMethodCall() {
+               nextToken();
+               return fToken == Symbols.TokenIDENT; // method name
+       }
+
+       /**
+        * Scans tokens for the matching opening peer. The internal cursor (<code>fPosition</code>)
+        * is set to the offset of the opening peer if found.
+        * 
+        * @return <code>true</code> if a matching token was found,
+        *         <code>false</code> otherwise
+        */
+       private boolean skipScope(int openToken, int closeToken) {
+
+               int depth = 1;
+
+               while (true) {
+                       nextToken();
+
+                       if (fToken == closeToken) {
+                               depth++;
+                       } else if (fToken == openToken) {
+                               depth--;
+                               if (depth == 0)
+                                       return true;
+                       } else if (fToken == Symbols.TokenEOF) {
+                               return false;
+                       }
+               }
+       }
+
+       // TODO adjust once there are per-project settings
+
+       private int prefTabLength() {
+               int tabLen;
+               // JavaCore core= JavaCore.getJavaCore();
+               WebUI plugin = WebUI.getDefault();
+               // if (core != null && plugin != null)
+               if (plugin != null)
+                       if (JavaCore.SPACE
+                                       .equals(JavaCore
+                                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_TAB_CHAR)))
+                               // if the formatter uses chars to mark indentation, then don't
+                               // substitute any chars
+                               tabLen = -1; // results in no tabs being substituted for
+                                                               // space runs
+                       else
+                               // if the formatter uses tabs to mark indentations, use the
+                               // visual setting from the editor
+                               // to get nicely aligned indentations
+                               tabLen = plugin
+                                               .getPreferenceStore()
+                                               .getInt(
+                                                               AbstractDecoratedTextEditorPreferenceConstants.EDITOR_TAB_WIDTH);
+               else
+                       tabLen = 4; // sensible default for testing
+
+               return tabLen;
+       }
+
+       private boolean prefArrayDimensionsDeepIndent() {
+               return true; // sensible default
+       }
+
+       private int prefArrayIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_EXPRESSIONS_IN_ARRAY_INITIALIZER);
+                       try {
+                               if (DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_BY_ONE)
+                                       return 1;
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+
+               return prefContinuationIndent(); // default
+       }
+
+       private boolean prefArrayDeepIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_EXPRESSIONS_IN_ARRAY_INITIALIZER);
+                       try {
+                               return DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_ON_COLUMN;
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+
+               return true;
+       }
+
+       private boolean prefTernaryDeepAlign() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_CONDITIONAL_EXPRESSION);
+                       try {
+                               return DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_ON_COLUMN;
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+               return false;
+       }
+
+       private int prefTernaryIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_CONDITIONAL_EXPRESSION);
+                       try {
+                               if (DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_BY_ONE)
+                                       return 1;
+                               else
+                                       return prefContinuationIndent();
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+
+               return prefContinuationIndent();
+       }
+
+       private int prefCaseIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       if (DefaultCodeFormatterConstants.TRUE
+                                       .equals(JavaCore
+                                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_INDENT_SWITCHSTATEMENTS_COMPARE_TO_SWITCH)))
+                               return prefBlockIndent();
+                       else
+                               return 0;
+               }
+
+               return 0; // sun standard
+       }
+
+       private int prefAssignmentIndent() {
+               return prefBlockIndent();
+       }
+
+       private int prefCaseBlockIndent() {
+               if (true)
+                       return prefBlockIndent();
+
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       if (DefaultCodeFormatterConstants.TRUE
+                                       .equals(JavaCore
+                                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_INDENT_SWITCHSTATEMENTS_COMPARE_TO_CASES)))
+                               return prefBlockIndent();
+                       else
+                               return 0;
+               }
+               return prefBlockIndent(); // sun standard
+       }
+
+       private int prefSimpleIndent() {
+               return prefBlockIndent();
+       }
+
+       private int prefBracketIndent() {
+               return prefBlockIndent();
+       }
+
+       private boolean prefMethodDeclDeepIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_PARAMETERS_IN_METHOD_DECLARATION);
+                       try {
+                               return DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_ON_COLUMN;
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+
+               return true;
+       }
+
+       private int prefMethodDeclIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_PARAMETERS_IN_METHOD_DECLARATION);
+                       try {
+                               if (DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_BY_ONE)
+                                       return 1;
+                               else
+                                       return prefContinuationIndent();
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+               return 1;
+       }
+
+       private boolean prefMethodCallDeepIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_ARGUMENTS_IN_METHOD_INVOCATION);
+                       try {
+                               return DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_ON_COLUMN;
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+               return false; // sensible default
+       }
+
+       private int prefMethodCallIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_ALIGNMENT_FOR_ARGUMENTS_IN_METHOD_INVOCATION);
+                       try {
+                               if (DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_BY_ONE)
+                                       return 1;
+                               else
+                                       return prefContinuationIndent();
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+
+               return 1; // sensible default
+       }
+
+       private boolean prefParenthesisDeepIndent() {
+
+               if (true) // don't do parenthesis deep indentation
+                       return false;
+
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_CONTINUATION_INDENTATION);
+                       try {
+                               return DefaultCodeFormatterConstants.getIndentStyle(option) == DefaultCodeFormatterConstants.INDENT_ON_COLUMN;
+                       } catch (IllegalArgumentException e) {
+                               // ignore and return default
+                       }
+               }
+
+               return false; // sensible default
+       }
+
+       private int prefParenthesisIndent() {
+               return prefContinuationIndent();
+       }
+
+       private int prefBlockIndent() {
+               return 1; // sensible default
+       }
+
+       private boolean prefIndentBracesForBlocks() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_BRACE_POSITION_FOR_BLOCK);
+                       return option
+                                       .equals(DefaultCodeFormatterConstants.NEXT_LINE_SHIFTED);
+               }
+
+               return false; // sensible default
+       }
+
+       private boolean prefIndentBracesForArrays() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_BRACE_POSITION_FOR_ARRAY_INITIALIZER);
+                       return option
+                                       .equals(DefaultCodeFormatterConstants.NEXT_LINE_SHIFTED);
+               }
+
+               return false; // sensible default
+       }
+
+       private boolean prefIndentBracesForMethods() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_BRACE_POSITION_FOR_METHOD_DECLARATION);
+                       return option
+                                       .equals(DefaultCodeFormatterConstants.NEXT_LINE_SHIFTED);
+               }
+
+               return false; // sensible default
+       }
+
+       private int prefContinuationIndent() {
+               Plugin plugin = JavaCore.getPlugin();
+               if (plugin != null) {
+                       String option = JavaCore
+                                       .getOption(DefaultCodeFormatterConstants.FORMATTER_CONTINUATION_INDENTATION);
+                       try {
+                               return Integer.parseInt(option);
+                       } catch (NumberFormatException e) {
+                               // ignore and return default
+                       }
+               }
+
+               return 2; // sensible default
+       }
+
+}