path: root/core/org.eclipse.cdt.core/src/org/eclipse/cdt/internal/formatter/Scribe.java

/*******************************************************************************
 * Copyright (c) 2000, 2016 IBM Corporation and others.
 *
 * This program and the accompanying materials
 * are made available under the terms of the Eclipse Public License 2.0
 * which accompanies this distribution, and is available at
 * https://www.eclipse.org/legal/epl-2.0/
 *
 * SPDX-License-Identifier: EPL-2.0
 *
 * Contributors:
 *     IBM Corporation - initial API and implementation
 *     Anton Leherbauer (Wind River Systems)
 *     Sergey Prigogin (Google)
 *******************************************************************************/
package org.eclipse.cdt.internal.formatter;

import java.util.Arrays;
import java.util.Collections;
import java.util.Iterator;
import java.util.List;

import org.eclipse.cdt.core.formatter.DefaultCodeFormatterConstants;
import org.eclipse.cdt.core.formatter.DefaultCodeFormatterOptions;
import org.eclipse.cdt.core.parser.util.CharArrayUtils;
import org.eclipse.cdt.internal.formatter.align.Alignment;
import org.eclipse.cdt.internal.formatter.align.AlignmentException;
import org.eclipse.cdt.internal.formatter.scanner.Scanner;
import org.eclipse.cdt.internal.formatter.scanner.Token;
import org.eclipse.jface.text.Position;
import org.eclipse.text.edits.MultiTextEdit;
import org.eclipse.text.edits.ReplaceEdit;
import org.eclipse.text.edits.TextEdit;

/**
 * This class is responsible for dumping formatted source.
 *
 * @since 4.0
 */
public class Scribe {
	private static final String EMPTY_STRING = ""; //$NON-NLS-1$
	private static final String SPACE = " "; //$NON-NLS-1$

	private static final int INITIAL_SIZE = 100;

	private final DefaultCodeFormatterOptions preferences;
	public final Scanner scanner;

	/** one-based column */
	public int column = 1;

	// Most specific alignment.
	public Alignment currentAlignment;
	public Alignment memberAlignment;
	public AlignmentException currentAlignmentException;

	/** @see Alignment#tailFormatter */
	private Runnable tailFormatter;

	public Token currentToken;

	// Edits management
	private OptimizedReplaceEdit[] edits;
	public int editsIndex;

	public int indentationLevel;
	public int numberOfIndentations;
	public int indentationSize;

	private final String lineSeparator;
	private final boolean indentEmptyLines;
	private final int pageWidth;
	private boolean preserveNewLines;
	private boolean checkLineWrapping;
	public int lastNumberOfNewLines;
	private boolean preserveLineBreakIndentation;
	boolean formatBrace;
	public int line;

	public boolean needSpace;
	public boolean pendingSpace;

	public int tabLength;
	public int tabChar;
	private final boolean useTabsOnlyForLeadingIndents;

	private int textRegionEnd;
	private int textRegionStart;
	public int scannerEndPosition;

	private List<Position> fSkipPositions = Collections.emptyList();

	private boolean skipOverInactive;

	private int fSkipStartOffset = Integer.MAX_VALUE;
	private int fSkipEndOffset;
	private int fSkippedIndentations;

	/* Comments formatting */
	private static final int NO_TRAILING_COMMENT = 0x0000;
	private static final int BASIC_TRAILING_COMMENT = 0x0100;
	private int[] lineOffsets;
	private int numLines;

	// Class to store previous line comment information
	static class LineComment {
		boolean contiguous;
		int currentIndentation;
		int indentation;
		int lines;
		char[] leadingSpaces = CharArrayUtils.EMPTY_CHAR_ARRAY;
	}

	final LineComment lastLineComment = new LineComment();

	Scribe(CodeFormatterVisitor formatter, int offset, int length) {
		scanner = new Scanner();
		preferences = formatter.preferences;
		pageWidth = preferences.page_width;
		tabLength = preferences.tab_size;
		indentationLevel = 0; // initialize properly
		numberOfIndentations = 0;
		useTabsOnlyForLeadingIndents = preferences.use_tabs_only_for_leading_indentations;
		indentEmptyLines = preferences.indent_empty_lines;
		tabChar = preferences.tab_char;
		if (tabChar == DefaultCodeFormatterOptions.MIXED) {
			indentationSize = preferences.indentation_size;
		} else {
			indentationSize = tabLength;
		}
		lineSeparator = preferences.line_separator;
		indentationLevel = preferences.initial_indentation_level * indentationSize;
		preserveNewLines = false;
		textRegionStart = offset;
		textRegionEnd = offset + length;
		reset();
	}

	private final void addDeleteEdit(int start, int end) {
		addOptimizedReplaceEdit(start, end - start + 1, EMPTY_STRING);
	}

	public final void addInsertEdit(int insertPosition, CharSequence insertedString) {
		addOptimizedReplaceEdit(insertPosition, 0, insertedString);
	}

	/**
	 * Adds a replace edit.
	 * @param start  start offset (inclusive)
	 * @param end  end offset (inclusive)
	 * @param replacement  the replacement string
	 */
	public final void addReplaceEdit(int start, int end, CharSequence replacement) {
		addOptimizedReplaceEdit(start, end - start + 1, replacement);
	}

	private final void addOptimizedReplaceEdit(int offset, int length, CharSequence replacement) {
		if (edits.length == editsIndex) {
			resize();
		}
		if (editsIndex > 0) {
			// Try to merge last two edits
			final OptimizedReplaceEdit previous = edits[editsIndex - 1];
			final int previousOffset = previous.offset;
			final int previousLength = previous.length;
			final int endOffsetOfPreviousEdit = previousOffset + previousLength;
			final int replacementLength = replacement.length();
			final String previousReplacement = previous.replacement;
			final int previousReplacementLength = previousReplacement.length();
			if (previousOffset == offset && previousLength == length
					&& (replacementLength == 0 || previousReplacementLength == 0)) {
				if (currentAlignment != null) {
					final Location location = currentAlignment.location;
					if (location.editsIndex == editsIndex) {
						location.editsIndex--;
						location.textEdit = previous;
					}
				}
				editsIndex--;
				return;
			}
			if (endOffsetOfPreviousEdit == offset) {
				if (length != 0) {
					if (replacementLength != 0) {
						editsIndex--;
						appendOptimizedReplaceEdit(previousOffset, previousLength + length,
								previousReplacement + replacement);
					} else if (previousLength + length == previousReplacementLength) {
						// Check the characters. If they are identical,
						// we can get rid of the previous edit.
						boolean canBeRemoved = true;
						loop: for (int i = previousOffset; i < previousOffset + previousReplacementLength; i++) {
							if (scanner.source[i] != previousReplacement.charAt(i - previousOffset)) {
								editsIndex--;
								appendOptimizedReplaceEdit(previousOffset, previousReplacementLength,
										previousReplacement);
								canBeRemoved = false;
								break loop;
							}
						}
						if (canBeRemoved) {
							if (currentAlignment != null) {
								final Location location = currentAlignment.location;
								if (location.editsIndex == editsIndex) {
									location.editsIndex--;
									location.textEdit = previous;
								}
							}
							editsIndex--;
						}
					} else {
						editsIndex--;
						appendOptimizedReplaceEdit(previousOffset, previousLength + length, previousReplacement);
					}
				} else {
					if (replacementLength != 0) {
						editsIndex--;
						appendOptimizedReplaceEdit(previousOffset, previousLength, previousReplacement + replacement);
					}
				}
			} else {
				assert endOffsetOfPreviousEdit < offset;
				appendOptimizedReplaceEdit(offset, length, replacement);
			}
		} else {
			appendOptimizedReplaceEdit(offset, length, replacement);
		}
	}

	/**
	 * Trims redundant prefix from a replacement edit and, if there is anything left, appends
	 * the replacement edit to the edits array.
	 */
	private void appendOptimizedReplaceEdit(int offset, int length, CharSequence replacement) {
		int replacementLength = replacement.length();
		int i;
		for (i = 0; i < length && i < replacementLength; i++, offset++) {
			if (scanner.source[offset] != replacement.charAt(i))
				break;
		}
		length -= i;
		if (i > 0) {
			replacement = i == replacementLength ? EMPTY_STRING : replacement.subSequence(i, replacementLength);
		}
		if (length > 0 || replacement.length() > 0) {
			edits[editsIndex++] = new OptimizedReplaceEdit(offset, length, replacement);
		}
	}

	public void alignFragment(Alignment alignment, int fragmentIndex) {
		alignment.alignFragment(fragmentIndex);
	}

	public void consumeNextToken() {
		printComment();
		currentToken = scanner.nextToken();
		addDeleteEdit(scanner.getCurrentTokenStartPosition(), scanner.getCurrentTokenEndPosition());
	}

	public Alignment createAlignment(String name, int mode, int count, int sourceRestart) {
		return createAlignment(name, mode, Alignment.R_INNERMOST, count, sourceRestart);
	}

	public Alignment createAlignment(String name, int mode, int count, int sourceRestart, boolean adjust) {
		return createAlignment(name, mode, Alignment.R_INNERMOST, count, sourceRestart, adjust);
	}

	public Alignment createAlignment(String name, int mode, int tieBreakRule, int count, int sourceRestart) {
		return createAlignment(name, mode, tieBreakRule, count, sourceRestart, preferences.continuation_indentation,
				false);
	}

	public Alignment createAlignment(String name, int mode, int count, int sourceRestart, int continuationIndent,
			boolean adjust) {
		return createAlignment(name, mode, Alignment.R_INNERMOST, count, sourceRestart, continuationIndent, adjust);
	}

	public Alignment createAlignment(String name, int mode, int tieBreakRule, int count, int sourceRestart,
			int continuationIndent, boolean adjust) {
		Alignment alignment = new Alignment(name, mode, tieBreakRule, this, count, sourceRestart, continuationIndent);
		// adjust break indentation
		if (adjust && memberAlignment != null) {
			Alignment current = memberAlignment;
			while (current.enclosing != null) {
				current = current.enclosing;
			}
			if ((current.mode & Alignment.M_MULTICOLUMN) != 0) {
				final int indentSize = indentationSize;
				switch (current.chunkKind) {
				case Alignment.CHUNK_METHOD:
				case Alignment.CHUNK_TYPE:
					if ((mode & Alignment.M_INDENT_BY_ONE) != 0) {
						alignment.breakIndentationLevel = indentationLevel + indentSize;
					} else {
						alignment.breakIndentationLevel = indentationLevel + continuationIndent * indentSize;
					}
					alignment.update();
					break;
				case Alignment.CHUNK_FIELD:
					if ((mode & Alignment.M_INDENT_BY_ONE) != 0) {
						alignment.breakIndentationLevel = current.originalIndentationLevel + indentSize;
					} else {
						alignment.breakIndentationLevel = current.originalIndentationLevel
								+ continuationIndent * indentSize;
					}
					alignment.update();
					break;
				}
			} else {
				switch (current.mode & Alignment.SPLIT_MASK) {
				case Alignment.M_COMPACT_SPLIT:
				case Alignment.M_COMPACT_FIRST_BREAK_SPLIT:
				case Alignment.M_NEXT_PER_LINE_SPLIT:
				case Alignment.M_NEXT_SHIFTED_SPLIT:
				case Alignment.M_ONE_PER_LINE_SPLIT:
					final int indentSize = indentationSize;
					switch (current.chunkKind) {
					case Alignment.CHUNK_METHOD:
					case Alignment.CHUNK_TYPE:
						if ((mode & Alignment.M_INDENT_BY_ONE) != 0) {
							alignment.breakIndentationLevel = indentationLevel + indentSize;
						} else {
							alignment.breakIndentationLevel = indentationLevel + continuationIndent * indentSize;
						}
						alignment.update();
						break;
					case Alignment.CHUNK_FIELD:
						if ((mode & Alignment.M_INDENT_BY_ONE) != 0) {
							alignment.breakIndentationLevel = current.originalIndentationLevel + indentSize;
						} else {
							alignment.breakIndentationLevel = current.originalIndentationLevel
									+ continuationIndent * indentSize;
						}
						alignment.update();
						break;
					}
					break;
				}
			}
		}
		return alignment;
	}

	public Alignment createMemberAlignment(String name, int mode, int count, int sourceRestart) {
		Alignment alignment = createAlignment(name, mode, Alignment.R_INNERMOST, count, sourceRestart);
		alignment.breakIndentationLevel = indentationLevel;
		return alignment;
	}

	public void enterAlignment(Alignment alignment) {
		alignment.enclosing = currentAlignment;
		currentAlignment = alignment;
	}

	public void enterMemberAlignment(Alignment alignment) {
		alignment.enclosing = memberAlignment;
		memberAlignment = alignment;
	}

	public void exitAlignment(Alignment alignment, boolean discardAlignment) {
		Alignment current = currentAlignment;
		while (current != null) {
			if (current == alignment)
				break;
			current = current.enclosing;
		}
		if (current == null) {
			throw new AbortFormatting("could not find matching alignment: " + alignment); //$NON-NLS-1$
		}
		indentationLevel = alignment.location.outputIndentationLevel;
		numberOfIndentations = alignment.location.numberOfIndentations;
		if (discardAlignment) {
			currentAlignment = alignment.enclosing;
		}
	}

	public void exitMemberAlignment(Alignment alignment) {
		Alignment current = memberAlignment;
		while (current != null) {
			if (current == alignment)
				break;
			current = current.enclosing;
		}
		if (current == null) {
			throw new AbortFormatting("could not find matching alignment: " + alignment); //$NON-NLS-1$
		}
		indentationLevel = current.location.outputIndentationLevel;
		numberOfIndentations = current.location.numberOfIndentations;
		memberAlignment = current.enclosing;
	}

	public Alignment getAlignment(String name) {
		if (currentAlignment != null) {
			return currentAlignment.getAlignment(name);
		}
		return null;
	}

	private int getIndentationOfOffset(int offset) {
		int beginningOfLine = getLineStart(offset);
		int indent = 0;
		for (int i = beginningOfLine; i < offset; i++) {
			indent = computeIndentation(scanner.source[i], indent);
		}
		return indent;
	}

	/**
	 * Computes indentation after applying a character at a given indentation position.
	 * @param text the text to be applied.
	 * @param indent the initial indentation.
	 * @return the resulting indentation.
	 */
	private int computeIndentation(char c, int indent) {
		switch (c) {
		case '\t':
			return tabLength > 0 ? indent + tabLength - indent % tabLength : indent;
		case '\r':
		case '\n':
			return 0;
		default:
			return indent + 1;
		}
	}

	/**
	 * Computes indentation after applying a given text at a given indentation position.
	 * @param text the text to be applied.
	 * @param indent the initial indentation.
	 * @return the resulting indentation.
	 */
	private int computeIndentation(char[] text, int indent) {
		if (text == null)
			return indent;
		int length = text.length;
		for (int i = 0; i < length; i++) {
			indent = computeIndentation(text[i], indent);
		}
		return indent;
	}

	private int computeIndentation(CharSequence text, int indent) {
		if (text == null)
			return indent;
		int length = text.length();
		for (int i = 0; i < length; i++) {
			indent = computeIndentation(text.charAt(i), indent);
		}
		return indent;
	}

	public String getEmptyLines(int linesNumber) {
		StringBuilder buffer = new StringBuilder();
		if (lastNumberOfNewLines == 0) {
			linesNumber++; // add an extra line breaks
			for (int i = 0; i < linesNumber; i++) {
				if (indentEmptyLines)
					printIndentationIfNecessary(buffer);
				buffer.append(lineSeparator);
			}
			lastNumberOfNewLines += linesNumber;
			line += linesNumber;
			column = 1;
			needSpace = false;
			pendingSpace = false;
		} else if (lastNumberOfNewLines == 1) {
			for (int i = 0; i < linesNumber; i++) {
				if (indentEmptyLines)
					printIndentationIfNecessary(buffer);
				buffer.append(lineSeparator);
			}
			lastNumberOfNewLines += linesNumber;
			line += linesNumber;
			column = 1;
			needSpace = false;
			pendingSpace = false;
		} else {
			if ((lastNumberOfNewLines - 1) >= linesNumber) {
				// there is no need to add new lines
				return EMPTY_STRING;
			}
			final int realNewLineNumber = linesNumber - lastNumberOfNewLines + 1;
			for (int i = 0; i < realNewLineNumber; i++) {
				if (indentEmptyLines)
					printIndentationIfNecessary(buffer);
				buffer.append(lineSeparator);
			}
			lastNumberOfNewLines += realNewLineNumber;
			line += realNewLineNumber;
			column = 1;
			needSpace = false;
			pendingSpace = false;
		}
		return String.valueOf(buffer);
	}

	public OptimizedReplaceEdit getLastEdit() {
		if (editsIndex > 0) {
			return edits[editsIndex - 1];
		}
		return null;
	}

	Alignment getMemberAlignment() {
		return memberAlignment;
	}

	public String getNewLine() {
		if (lastNumberOfNewLines >= 1) {
			column = 1; // Ensure that the scribe is at the beginning of a new line
			return EMPTY_STRING;
		}
		line++;
		lastNumberOfNewLines = 1;
		column = 1;
		needSpace = false;
		pendingSpace = false;
		return lineSeparator;
	}

	/**
	 * Returns next indentation level based on column estimated position (if
	 * column is not indented, then uses indentationLevel)
	 */
	public int getNextIndentationLevel(int someColumn) {
		int indent = someColumn - 1;
		if (indent == 0)
			return indentationLevel;
		if (tabChar == DefaultCodeFormatterOptions.TAB && !useTabsOnlyForLeadingIndents) {
			// Round up to a multiple of indentationSize.
			indent += indentationSize - 1;
			return indent - indent % indentationSize;
		} else {
			return indent;
		}
	}

	private String getPreserveEmptyLines(int count) {
		if (count == 0 && !preserveNewLines) {
			// preserve line breaks in wrapping if specified
			if (!preferences.join_wrapped_lines && lastNumberOfNewLines == 0) {
				// Create new line
				StringBuilder tempBuffer = new StringBuilder();
				tempBuffer.append(getNewLine());

				if (currentAlignment != null && !formatBrace) {
					indentationLevel = currentAlignment.breakIndentationLevel;
				}

				// Set the flag to indicate that a specific indentation is currently in used
				preserveLineBreakIndentation = true;

				// Print the computed indentation in the buffer
				printIndentationIfNecessary(tempBuffer);

				return tempBuffer.toString();
			}
			return EMPTY_STRING;
		}
		if (preferences.number_of_empty_lines_to_preserve != 0) {
			int linesToPreserve = Math.min(count, preferences.number_of_empty_lines_to_preserve);
			return getEmptyLines(linesToPreserve);
		} else if (preserveNewLines) {
			return getNewLine();
		}
		return EMPTY_STRING;
	}

	public TextEdit getRootEdit() {
		MultiTextEdit edit = null;
		int length = textRegionEnd - textRegionStart;
		if (textRegionStart <= 0) {
			if (length <= 0) {
				edit = new MultiTextEdit(0, 0);
			} else {
				edit = new MultiTextEdit(0, textRegionEnd);
			}
		} else {
			edit = new MultiTextEdit(textRegionStart, length);
		}
		for (int i = 0, max = editsIndex; i < max; i++) {
			OptimizedReplaceEdit currentEdit = edits[i];
			if (isValidEdit(currentEdit)) {
				edit.addChild(new ReplaceEdit(currentEdit.offset, currentEdit.length, currentEdit.replacement));
			}
		}
		edits = null;
		return edit;
	}

	public void handleLineTooLong() {
		// Search for closest breakable alignment, using tie-breaking rules.
		// Look for innermost breakable one.
		int relativeDepth = 0;
		Alignment targetAlignment = currentAlignment;
		while (targetAlignment != null && targetAlignment.tieBreakRule == Alignment.R_INNERMOST) {
			if (targetAlignment.couldBreak()) {
				throwAlignmentException(AlignmentException.LINE_TOO_LONG, relativeDepth);
			}
			targetAlignment = targetAlignment.enclosing;
			relativeDepth++;
		}

		// Look for outermost breakable one.
		relativeDepth = 0;
		int outerMostDepth = -1;
		targetAlignment = currentAlignment;
		while (targetAlignment != null) {
			if (targetAlignment.tieBreakRule == Alignment.R_OUTERMOST && targetAlignment.couldBreak()) {
				outerMostDepth = relativeDepth;
			}
			targetAlignment = targetAlignment.enclosing;
			relativeDepth++;
		}
		if (outerMostDepth >= 0) {
			throwAlignmentException(AlignmentException.LINE_TOO_LONG, outerMostDepth);
		}

		// Look for innermost breakable one but don't stop if we encounter a R_OUTERMOST
		// tie-breaking rule.
		relativeDepth = 0;
		targetAlignment = currentAlignment;
		while (targetAlignment != null) {
			if (targetAlignment.couldBreak()) {
				throwAlignmentException(AlignmentException.LINE_TOO_LONG, relativeDepth);
			}
			targetAlignment = targetAlignment.enclosing;
			relativeDepth++;
		}
		// Did not find any breakable location - proceed
	}

	private void throwAlignmentException(int kind, int relativeDepth) {
		AlignmentException e = new AlignmentException(kind, relativeDepth);
		currentAlignmentException = e;
		throw e;
	}

	public void indent() {
		if (shouldSkip(scanner.getCurrentPosition())) {
			fSkippedIndentations++;
			return;
		}
		indentationLevel += indentationSize;
		numberOfIndentations++;
	}

	/**
	 * @param translationUnitSource
	 */
	public void initializeScanner(char[] translationUnitSource) {
		scanner.setSource(translationUnitSource);
		scannerEndPosition = translationUnitSource.length;
		scanner.resetTo(0, scannerEndPosition);
		edits = new OptimizedReplaceEdit[INITIAL_SIZE];
		// Locate line breaks.
		lineOffsets = new int[200];
		numLines = 0;
		lineOffsets[numLines++] = 0;
		for (int i = 0; i < translationUnitSource.length; i++) {
			if (translationUnitSource[i] == '\n') {
				int len = lineOffsets.length;
				if (numLines >= len)
					System.arraycopy(lineOffsets, 0, lineOffsets = new int[len + (len + 1) / 2], 0, len);
				lineOffsets[numLines++] = i + 1;
			}
		}
	}

	/**
	 * @param list
	 */
	public void setSkipPositions(List<Position> list) {
		fSkipPositions = list;
		skipOverInactive = !list.isEmpty();
	}

	/**
	 * Returns offset of the start of the line containing a given offset.
	 */
	private int getLineStart(int offset) {
		// Binary search
		int down = 0;
		int up = numLines;
		while (true) {
			int mid = (down + up) / 2;
			int lineOffset = lineOffsets[mid];
			if (mid == down) {
				return lineOffset;
			}
			if (lineOffset > offset) {
				up = mid;
			} else {
				down = mid;
			}
		}
	}

	private boolean isOnFirstColumn(int offset) {
		return getLineStart(offset) == offset;
	}

	private boolean isValidEdit(OptimizedReplaceEdit edit) {
		final int editLength = edit.length;
		final int editReplacementLength = edit.replacement.length();
		final int editOffset = edit.offset;
		if (editLength != 0) {
			if (textRegionStart <= editOffset && editOffset + editLength <= textRegionEnd) {
				if (editReplacementLength != 0 && editLength == editReplacementLength) {
					for (int i = editOffset, max = editOffset + editLength; i < max; i++) {
						if (scanner.source[i] != edit.replacement.charAt(i - editOffset)) {
							return true;
						}
					}
					return false;
				} else {
					return true;
				}
			} else if (editOffset + editLength == textRegionStart) {
				int i = editOffset;
				for (int max = editOffset + editLength; i < max; i++) {
					int replacementStringIndex = i - editOffset;
					if (replacementStringIndex >= editReplacementLength
							|| scanner.source[i] != edit.replacement.charAt(replacementStringIndex)) {
						break;
					}
				}
				if (i - editOffset != editReplacementLength && i != editOffset + editLength - 1) {
					edit.offset = textRegionStart;
					edit.length = 0;
					edit.replacement = edit.replacement.substring(i - editOffset);
					return true;
				}
			}
		} else if (textRegionStart <= editOffset && editOffset < textRegionEnd) {
			return true;
		} else if (editOffset == scannerEndPosition && editOffset == textRegionEnd) {
			return true;
		}
		return false;
	}

	private void preserveEmptyLines(int count, int insertPosition) {
		if (count > 0) {
			if (preferences.number_of_empty_lines_to_preserve != 0) {
				int linesToPreserve = Math.min(count, preferences.number_of_empty_lines_to_preserve);
				printEmptyLines(linesToPreserve, insertPosition);
			} else {
				printNewLine(insertPosition);
			}
		} else {
			printNewLine(insertPosition);
		}
	}

	public void printRaw(int startOffset, int length) {
		if (length <= 0) {
			return;
		}
		int currentPosition = scanner.getCurrentPosition();
		if (shouldSkip(currentPosition)) {
			return;
		}
		if (startOffset > currentPosition) {
			printComment();
			currentPosition = scanner.getCurrentPosition();
		}
		if (startOffset + length < currentPosition) {
			return; // Don't move backwards
		}
		boolean savedPreserveNL = preserveNewLines;
		boolean savedSkipOverInactive = skipOverInactive;
		int savedScannerEndPos = scannerEndPosition;
		preserveNewLines = true;
		skipOverInactive = false;
		scannerEndPosition = startOffset + length;
		try {
			scanner.resetTo(Math.max(startOffset, currentPosition), startOffset + length);
			int parenLevel = 0;
			while (true) {
				boolean hasWhitespace = printComment();
				currentToken = scanner.nextToken();
				if (currentToken == null) {
					if (hasWhitespace) {
						space();
					}
					break;
				}
				if (pendingSpace) {
					addInsertEdit(scanner.getCurrentTokenStartPosition(), SPACE);
					pendingSpace = false;
					needSpace = false;
				}
				switch (currentToken.type) {
				case Token.tLBRACE: {
					scanner.resetTo(scanner.getCurrentTokenStartPosition(), scannerEndPosition);
					formatOpeningBrace(preferences.brace_position_for_block,
							preferences.insert_space_before_opening_brace_in_block);
					if (preferences.indent_statements_compare_to_block) {
						indent();
					}
					break;
				}
				case Token.tRBRACE: {
					scanner.resetTo(scanner.getCurrentTokenStartPosition(), scannerEndPosition);
					if (preferences.indent_statements_compare_to_block) {
						unIndent();
					}
					formatClosingBrace(preferences.brace_position_for_block);
					break;
				}
				case Token.tLPAREN:
					++parenLevel;
					print(currentToken.getLength(), hasWhitespace);
					if (parenLevel > 0) {
						indentForContinuation();
						if (column <= indentationLevel) {
							// HACK: avoid indent in same line
							column = indentationLevel + 1;
						}
					}
					break;
				case Token.tRPAREN:
					--parenLevel;
					if (parenLevel >= 0) {
						unIndentForContinuation();
					}
					print(currentToken.getLength(), hasWhitespace);
					break;
				case Token.tSEMI:
					print(currentToken.getLength(), preferences.insert_space_before_semicolon);
					break;
				case Token.t_catch:
				case Token.t_else:
					if (preferences.insert_new_line_before_else_in_if_statement) {
						printNewLine(currentToken.offset);
					} else {
						hasWhitespace = true;
					}
					print(currentToken.getLength(), hasWhitespace);
					break;
				default:
					if (currentToken.isVisibilityModifier()
							&& !preferences.indent_access_specifier_compare_to_type_header) {
						int indentLevel = indentationLevel;
						if (indentationLevel > 0)
							unIndent();
						print(currentToken.getLength(), hasWhitespace);
						while (indentationLevel < indentLevel) {
							indent();
						}
					} else {
						print(currentToken.getLength(), hasWhitespace);
					}
				}
				hasWhitespace = false;
			}
		} finally {
			scannerEndPosition = savedScannerEndPos;
			scanner.resetTo(startOffset + length, scannerEndPosition);
			skipOverInactive = savedSkipOverInactive;
			preserveNewLines = savedPreserveNL;
		}
	}

	public void indentForContinuation() {
		for (int i = 0; i < preferences.continuation_indentation; i++) {
			indent();
		}
	}

	public void unIndentForContinuation() {
		for (int i = 0; i < preferences.continuation_indentation; i++) {
			unIndent();
		}
	}

	public void formatOpeningBrace(String bracePosition, boolean insertSpaceBeforeBrace) {
		if (DefaultCodeFormatterConstants.NEXT_LINE.equals(bracePosition)) {
			printNewLine();
		} else if (DefaultCodeFormatterConstants.NEXT_LINE_SHIFTED.equals(bracePosition)) {
			printNewLine();
			indent();
		}
		printNextToken(Token.tLBRACE, insertSpaceBeforeBrace);

		printTrailingComment();
	}

	public void formatClosingBrace(String block_brace_position) {
		printNextToken(Token.tRBRACE);
		printTrailingComment();
		if (DefaultCodeFormatterConstants.NEXT_LINE_SHIFTED.equals(block_brace_position)) {
			unIndent();
		}
	}

	private void print(int length, boolean considerSpaceIfAny) {
		if (checkLineWrapping && length + column - 1 > pageWidth) {
			handleLineTooLong();
		}
		lastNumberOfNewLines = 0;
		if (indentationLevel != 0) {
			printIndentationIfNecessary();
		}
		if (considerSpaceIfAny) {
			space();
		}
		if (pendingSpace) {
			addInsertEdit(scanner.getCurrentTokenStartPosition(), SPACE);
		}
		if (checkLineWrapping && length + column - 1 > pageWidth) {
			handleLineTooLong();
		}
		pendingSpace = false;
		column += length;
		needSpace = true;
	}

	private void printBlockComment(boolean forceNewLine) {
		int currentTokenStartPosition = scanner.getCurrentTokenStartPosition();
		int currentTokenEndPosition = scanner.getCurrentTokenEndPosition() + 1;

		scanner.resetTo(currentTokenStartPosition, currentTokenEndPosition);
		int currentCharacter;
		boolean isNewLine = false;
		int start = currentTokenStartPosition;
		int nextCharacterStart = currentTokenStartPosition;
		printIndentationIfNecessary();
		if (pendingSpace) {
			addInsertEdit(currentTokenStartPosition, SPACE);
		}
		needSpace = false;
		pendingSpace = false;
		int previousStart = currentTokenStartPosition;

		while (nextCharacterStart <= currentTokenEndPosition && (currentCharacter = scanner.getNextChar()) != -1) {
			nextCharacterStart = scanner.getCurrentPosition();

			switch (currentCharacter) {
			case '\r':
				if (isNewLine) {
					line++;
				}
				start = previousStart;
				isNewLine = true;
				if (scanner.getNextChar('\n')) {
					currentCharacter = '\n';
					nextCharacterStart = scanner.getCurrentPosition();
				}
				break;
			case '\n':
				if (isNewLine) {
					line++;
				}
				start = previousStart;
				isNewLine = true;
				break;
			default:
				if (isNewLine) {
					if (Character.isWhitespace((char) currentCharacter)) {
						int previousStartPosition = scanner.getCurrentPosition();
						while (currentCharacter != -1 && currentCharacter != '\r' && currentCharacter != '\n'
								&& Character.isWhitespace((char) currentCharacter)) {
							previousStart = nextCharacterStart;
							previousStartPosition = scanner.getCurrentPosition();
							currentCharacter = scanner.getNextChar();
							nextCharacterStart = scanner.getCurrentPosition();
						}
						if (currentCharacter == '\r' || currentCharacter == '\n') {
							nextCharacterStart = previousStartPosition;
						}
					}
					column = 1;
					line++;

					StringBuilder buffer = new StringBuilder();
					buffer.append(lineSeparator);
					printIndentationIfNecessary(buffer);
					buffer.append(' ');

					addReplaceEdit(start, previousStart - 1, String.valueOf(buffer));
				} else {
					column += (nextCharacterStart - previousStart);
				}
				isNewLine = false;
			}
			previousStart = nextCharacterStart;
			scanner.setCurrentPosition(nextCharacterStart);
		}
		lastNumberOfNewLines = 0;
		needSpace = false;
		scanner.resetTo(currentTokenEndPosition, scannerEndPosition);
		if (forceNewLine) {
			startNewLine();
		}
	}

	private void printPreprocessorDirective() {
		int currentTokenStartPosition = scanner.getCurrentTokenStartPosition();
		int currentTokenEndPosition = scanner.getCurrentTokenEndPosition() + 1;

		scanner.resetTo(currentTokenStartPosition, currentTokenEndPosition);
		int currentCharacter;
		boolean isNewLine = false;
		int nextCharacterStart = currentTokenStartPosition;
		needSpace = false;
		pendingSpace = false;
		int previousStart = currentTokenStartPosition;

		while (nextCharacterStart <= currentTokenEndPosition && (currentCharacter = scanner.getNextChar()) != -1) {
			nextCharacterStart = scanner.getCurrentPosition();

			switch (currentCharacter) {
			case '\r':
				isNewLine = true;
				if (scanner.getNextChar('\n')) {
					currentCharacter = '\n';
					nextCharacterStart = scanner.getCurrentPosition();
				}
				break;
			case '\n':
				isNewLine = true;
				break;
			default:
				if (isNewLine) {
					column = 1;
					line++;
				} else {
					column += (nextCharacterStart - previousStart);
				}
				isNewLine = false;
			}
			previousStart = nextCharacterStart;
			scanner.setCurrentPosition(nextCharacterStart);
		}
		lastNumberOfNewLines = isNewLine ? 1 : 0;
		needSpace = false;
		if (currentAlignment != null) {
			indentationLevel = currentAlignment.breakIndentationLevel;
		}
		scanner.resetTo(currentTokenEndPosition, scannerEndPosition);
	}

	public void printEndOfTranslationUnit() {
		int currentTokenStartPosition = scanner.getCurrentPosition();
		if (currentTokenStartPosition <= scannerEndPosition) {
			printRaw(currentTokenStartPosition, scannerEndPosition - currentTokenStartPosition + 1);
		}
	}

	public boolean printComment() {
		return printComment(NO_TRAILING_COMMENT);
	}

	/**
	 * Prints comment at the current position.
	 *
	 * @return {@code true} if a writespace character was encountered preceding the next token,
	 */
	public boolean printComment(int trailing) {
		// If we have a space between two tokens we ensure it will be dumped in the formatted
		// string.
		int currentTokenStartPosition = scanner.getCurrentPosition();
		if (shouldSkip(currentTokenStartPosition)) {
			return false;
		}
		boolean hasComment = false;
		boolean hasLineComment = false;
		boolean hasWhitespace = false;
		char[] whiteSpaces = CharArrayUtils.EMPTY_CHAR_ARRAY;
		int lines = 0;
		while ((currentToken = scanner.nextToken()) != null) {
			if (skipOverInactive) {
				Position inactivePos = getInactivePosAt(scanner.getCurrentTokenStartPosition());
				if (inactivePos != null) {
					int startOffset = Math.min(scanner.getCurrentTokenStartPosition(), inactivePos.getOffset());
					int endOffset = Math.min(scannerEndPosition, inactivePos.getOffset() + inactivePos.getLength());
					if (startOffset < endOffset) {
						int savedIndentLevel = indentationLevel;
						scanner.resetTo(scanner.getCurrentTokenStartPosition(), scanner.eofPosition);
						printRaw(startOffset, endOffset - startOffset);
						while (indentationLevel > savedIndentLevel) {
							unIndent();
						}
						while (indentationLevel < savedIndentLevel) {
							indent();
						}
						scanner.resetTo(endOffset, scanner.eofPosition);
						continue;
					}
				}
			}
			int tokenStartPosition = scanner.getCurrentTokenStartPosition();
			switch (currentToken.type) {
			case Token.tWHITESPACE:
				whiteSpaces = scanner.getCurrentTokenSource();
				int whitespacesEndPosition = scanner.getCurrentTokenEndPosition();
				lines = 0;
				for (int i = 0, max = whiteSpaces.length; i < max; i++) {
					switch (whiteSpaces[i]) {
					case '\r':
						if ((i + 1) < max) {
							if (whiteSpaces[i + 1] == '\n') {
								i++;
							}
						}
						lines++;
						break;
					case '\n':
						lines++;
					}
				}
				// If following token is a line comment on the same line or the line just after,
				// then it might be not really formatted as a trailing comment.
				boolean realTrailing = trailing != NO_TRAILING_COMMENT;
				if (realTrailing && scanner.peekNextChar() == '/' && lines == 0) {
					boolean canChangeTrailing = false;
					// For basic trailing comment preceded by a line comment, then it depends on
					// the comments relative position when following comment column (after having
					// been rounded) is below the preceding one, then it becomes not a good idea
					// to change the trailing flag.
					if (trailing == BASIC_TRAILING_COMMENT && hasLineComment) {
						int currentCommentIndentation = computeIndentation(whiteSpaces, 0);
						int relativeIndentation = currentCommentIndentation - lastLineComment.currentIndentation;
						if (tabLength == 0) {
							canChangeTrailing = relativeIndentation == 0;
						} else {
							canChangeTrailing = relativeIndentation > -tabLength;
						}
					}
					// If the trailing can be changed, then look at the following tokens.
					if (canChangeTrailing) {
						int currentTokenPosition = scanner.getCurrentTokenStartPosition();
						if (scanner.getNextToken() == Token.tLINECOMMENT) {
							realTrailing = !hasLineComment;
							switch (scanner.getNextToken()) {
							case Token.tLINECOMMENT:
								// At least two contiguous line comments.
								// The formatter should not consider comments as trailing ones.
								realTrailing = false;
								break;
							case Token.tWHITESPACE:
								if (scanner.getNextToken() == Token.tLINECOMMENT) {
									// At least two contiguous line comments.
									// The formatter should not consider comments as trailing ones.
									realTrailing = false;
								}
								break;
							}
						}
						scanner.resetTo(currentTokenPosition, scanner.eofPosition);
						scanner.getNextToken(); // Get current token again to restore the scanner state.
					}
				}
				// Strategy to consume spaces and eventually leave at this stage
				// depends on the fact that a trailing comment is expected or not
				if (realTrailing) {
					// If a line comment is consumed, no other comment can be on the same line after
					if (hasLineComment) {
						if (lines >= 1) {
							currentTokenStartPosition = tokenStartPosition;
							preserveEmptyLines(lines, currentTokenStartPosition);
							addDeleteEdit(currentTokenStartPosition, whitespacesEndPosition);
							scanner.resetTo(scanner.getCurrentPosition(), scannerEndPosition);
							return hasWhitespace;
						}
						scanner.resetTo(currentTokenStartPosition, scannerEndPosition);
						return hasWhitespace;
					}
					// If one or several new lines are consumed, following comments
					// cannot be considered as trailing ones.
					if (lines >= 1) {
						if (hasComment) {
							printNewLine(tokenStartPosition);
						}
						scanner.resetTo(currentTokenStartPosition, scannerEndPosition);
						return hasWhitespace;
					}
					// Delete consumed white spaces
					hasWhitespace = true;
					currentTokenStartPosition = scanner.getCurrentPosition();
					addDeleteEdit(tokenStartPosition, whitespacesEndPosition);
				} else {
					if (lines == 0) {
						hasWhitespace = true;
						addDeleteEdit(scanner.getCurrentTokenStartPosition(), scanner.getCurrentTokenEndPosition());
					} else if (hasComment) {
						if (lines == 1) {
							printNewLine(scanner.getCurrentTokenStartPosition());
						} else {
							preserveEmptyLines(lines - 1, scanner.getCurrentTokenStartPosition());
						}
						addDeleteEdit(scanner.getCurrentTokenStartPosition(), scanner.getCurrentTokenEndPosition());
					} else if (hasLineComment) {
						preserveEmptyLines(lines, scanner.getCurrentTokenStartPosition());
						addDeleteEdit(scanner.getCurrentTokenStartPosition(), scanner.getCurrentTokenEndPosition());
					} else if (lines != 0 && (!preferences.join_wrapped_lines
							|| preferences.number_of_empty_lines_to_preserve != 0)) {
						String preservedEmptyLines = getPreserveEmptyLines(lines - 1);
						addReplaceEdit(scanner.getCurrentTokenStartPosition(), scanner.getCurrentTokenEndPosition(),
								preservedEmptyLines);
						hasWhitespace = preservedEmptyLines.length() == 0;
					} else {
						addDeleteEdit(scanner.getCurrentTokenStartPosition(), scanner.getCurrentTokenEndPosition());
						hasWhitespace = true;
					}
				}
				currentTokenStartPosition = scanner.getCurrentPosition();
				break;
			case Token.tLINECOMMENT:
				if (lines >= 1) {
					if (lines > 1) {
						preserveEmptyLines(lines - 1, scanner.getCurrentTokenStartPosition());
					} else if (lines == 1) {
						printNewLine(scanner.getCurrentTokenStartPosition());
					}
				} else if (hasWhitespace) {
					// Look whether comments line may be contiguous or not
					// Note that when preceding token is a comment line, then only one line
					// is enough to have an empty line as the line end is included in the comment line.
					// If comments are contiguous, store the white spaces to be able to compute
					// the current comment indentation
					if (lines > 1 || (lines == 1 && hasLineComment)) {
						lastLineComment.contiguous = false;
					}
					lastLineComment.leadingSpaces = whiteSpaces;
					lastLineComment.lines = lines;
				}
				whiteSpaces = CharArrayUtils.EMPTY_CHAR_ARRAY;
				hasWhitespace = false;
				printLineComment();
				currentTokenStartPosition = scanner.getCurrentPosition();
				hasLineComment = true;
				lines = 0;
				break;
			case Token.tBLOCKCOMMENT:
				if (lines >= 1) {
					if (lines > 1) {
						preserveEmptyLines(lines - 1, scanner.getCurrentTokenStartPosition());
					} else if (lines == 1) {
						printNewLine(scanner.getCurrentTokenStartPosition());
					}
				} else if (hasWhitespace) {
					space();
				}
				whiteSpaces = CharArrayUtils.EMPTY_CHAR_ARRAY;
				hasWhitespace = false;
				printBlockComment(false);
				currentTokenStartPosition = scanner.getCurrentPosition();
				hasLineComment = false;
				hasComment = true;
				lines = 0;
				break;
			case Token.tPREPROCESSOR:
			case Token.tPREPROCESSOR_DEFINE:
			case Token.tPREPROCESSOR_INCLUDE:
				if (column != 1)
					printNewLine(scanner.getCurrentTokenStartPosition());
				if (lines >= 1) {
					if (lines > 1) {
						preserveEmptyLines(lines - 1, scanner.getCurrentTokenStartPosition());
					} else if (lines == 1) {
						// printNewLine(scanner.getCurrentTokenStartPosition());
					}
				}
				whiteSpaces = CharArrayUtils.EMPTY_CHAR_ARRAY;
				hasWhitespace = false;
				printPreprocessorDirective();
				printNewLine();
				currentTokenStartPosition = scanner.getCurrentPosition();
				hasLineComment = false;
				hasComment = false;
				lines = 0;
				break;
			default:
				// step back one token
				scanner.resetTo(currentTokenStartPosition, scannerEndPosition);
				return hasWhitespace;
			}
		}
		scanner.resetTo(currentTokenStartPosition, scannerEndPosition);
		return hasWhitespace;
	}

	/**
	 * @param offset
	 * @return
	 */
	private Position getInactivePosAt(int offset) {
		for (Iterator<Position> iter = fSkipPositions.iterator(); iter.hasNext();) {
			Position pos = iter.next();
			if (pos.includes(offset)) {
				return pos;
			}
		}
		return null;
	}

	private void printLineComment() {
		int currentTokenStartPosition = scanner.getCurrentTokenStartPosition();
		int currentTokenEndPosition = scanner.getCurrentTokenEndPosition() + 1;
		scanner.resetTo(currentTokenStartPosition, currentTokenEndPosition);
		int currentCharacter;
		int start = currentTokenStartPosition;
		int nextCharacterStart = currentTokenStartPosition;

		// Print comment line indentation
		int commentIndentationLevel;
		boolean onFirstColumn = isOnFirstColumn(start);
		if (!preferences.comment_line_up_line_comment_in_blocks_on_first_column && indentationLevel == 0) {
			commentIndentationLevel = column - 1;
		} else {
			if (onFirstColumn && preferences.never_indent_line_comments_on_first_column) {
				commentIndentationLevel = column - 1;
			} else {
				// Indentation may be specific for contiguous comment
				// see bug https://bugs.eclipse.org/bugs/show_bug.cgi?id=293300
				if (lastLineComment.contiguous) {
					// The leading spaces have been set while looping in the printComment(int) method
					int currentCommentIndentation = computeIndentation(lastLineComment.leadingSpaces, 0);
					// Keep the current comment indentation when over the previous contiguous line comment
					// and the previous comment has not been re-indented
					int relativeIndentation = currentCommentIndentation - lastLineComment.currentIndentation;
					boolean similarCommentsIndentation = false;
					if (tabLength == 0) {
						similarCommentsIndentation = relativeIndentation == 0;
					} else if (relativeIndentation > -tabLength) {
						similarCommentsIndentation = relativeIndentation == 0
								|| currentCommentIndentation != 0 && lastLineComment.currentIndentation != 0;
					}
					if (similarCommentsIndentation && lastLineComment.indentation != indentationLevel) {
						int currentIndentationLevel = indentationLevel;
						indentationLevel = lastLineComment.indentation;
						printIndentationIfNecessary();
						indentationLevel = currentIndentationLevel;
						commentIndentationLevel = lastLineComment.indentation;
					} else {
						printIndentationIfNecessary();
						commentIndentationLevel = column - 1;
					}
				} else {
					printIndentationIfNecessary();
					commentIndentationLevel = column - 1;
				}
			}
		}

		// Prepare white space before the comment.
		StringBuilder whitespace = null;
		if (!lastLineComment.contiguous && commentIndentationLevel != indentationLevel
				&& !(preferences.never_indent_line_comments_on_first_column && onFirstColumn)) {
			whitespace = new StringBuilder();
			int whitespaceStartPosition = currentTokenStartPosition - lastLineComment.leadingSpaces.length;
			int indent = getIndentationOfOffset(whitespaceStartPosition);
			int distance = computeIndentation(lastLineComment.leadingSpaces, indent) - indent;
			if (preferences.comment_preserve_white_space_between_code_and_line_comment
					&& distance >= preferences.comment_min_distance_between_code_and_line_comment) {
				whitespace.append(lastLineComment.leadingSpaces);
			} else {
				for (int i = 0; i < preferences.comment_min_distance_between_code_and_line_comment; i++) {
					whitespace.append(' ');
				}
			}
		}

		// Store line comment information
		lastLineComment.currentIndentation = getIndentationOfOffset(currentTokenStartPosition);
		lastLineComment.contiguous = true;

		while (true) {
			Location location = new Location(this, scanner.getCurrentPosition());
			int commentIndent = commentIndentationLevel;

			// Add pending space if necessary
			if (whitespace != null) {
				addInsertEdit(currentTokenStartPosition, whitespace);
				commentIndent = computeIndentation(whitespace, commentIndentationLevel);
				needSpace = false;
				pendingSpace = false;
			}
			lastLineComment.indentation = commentIndent;

			int previousStart = currentTokenStartPosition;

			int indent = commentIndent;
			loop: while (nextCharacterStart <= currentTokenEndPosition
					&& (currentCharacter = scanner.getNextChar()) != -1) {
				nextCharacterStart = scanner.getCurrentPosition();

				switch (currentCharacter) {
				case '\r':
				case '\n':
					start = previousStart;
					break loop;
				}
				indent = computeIndentation((char) currentCharacter, indent);
				previousStart = nextCharacterStart;
			}

			if (start != currentTokenStartPosition) {
				// This means that the line comment doesn't end the file
				addReplaceEdit(start, currentTokenEndPosition - 1, lineSeparator);
				line++;
				column = 1;
				lastNumberOfNewLines = 1;
			}
			if (!checkLineWrapping || indent <= pageWidth || whitespace == null || commentIndent
					- commentIndentationLevel <= preferences.comment_min_distance_between_code_and_line_comment) {
				break;
			}

			// Maximum line length was exceeded. Try to reduce white space before the comment by
			// removing the last white space character.
			whitespace.deleteCharAt(whitespace.length() - 1);
			if (computeIndentation(lastLineComment.leadingSpaces, commentIndentationLevel)
					- commentIndentationLevel < preferences.comment_min_distance_between_code_and_line_comment) {
				// The white space shrank too much. Rebuild it to satisfy the minimum distance
				// requirement.
				whitespace.delete(0, whitespace.length());
				for (int i = 0; i < preferences.comment_min_distance_between_code_and_line_comment; i++) {
					whitespace.append(' ');
				}
			}
			resetAt(location);
			scanner.resetTo(location.inputOffset, scanner.eofPosition);
		}

		needSpace = false;
		pendingSpace = false;
		// realign to the proper value
		if (currentAlignment != null) {
			if (memberAlignment != null) {
				// select the last alignment
				if (currentAlignment.location.inputOffset > memberAlignment.location.inputOffset) {
					if (currentAlignment.couldBreak() && currentAlignment.wasSplit) {
						currentAlignment.performFragmentEffect();
					}
				} else {
					indentationLevel = Math.max(indentationLevel, memberAlignment.breakIndentationLevel);
				}
			} else if (currentAlignment.couldBreak() && currentAlignment.wasSplit) {
				currentAlignment.performFragmentEffect();
			}
			if (currentAlignment.name.equals(Alignment.BINARY_EXPRESSION) && currentAlignment.enclosing != null
					&& currentAlignment.enclosing.equals(Alignment.BINARY_EXPRESSION)
					&& indentationLevel < currentAlignment.breakIndentationLevel) {
				indentationLevel = currentAlignment.breakIndentationLevel;
			}
		}
		scanner.resetTo(currentTokenEndPosition, scannerEndPosition);
	}

	public void printEmptyLines(int linesNumber) {
		printEmptyLines(linesNumber, scanner.getCurrentTokenEndPosition() + 1);
	}

	private void printEmptyLines(int linesNumber, int insertPosition) {
		final String buffer = getEmptyLines(linesNumber);
		if (!buffer.isEmpty()) {
			addInsertEdit(insertPosition, buffer);
		}
	}

	void printIndentationIfNecessary() {
		if (column > indentationLevel)
			return;
		StringBuilder buffer = new StringBuilder();
		printIndentationIfNecessary(buffer);
		if (buffer.length() > 0) {
			addInsertEdit(scanner.getCurrentTokenStartPosition(), buffer);
			pendingSpace = false;
		}
	}

	private void printIndentationIfNecessary(StringBuilder buffer) {
		switch (tabChar) {
		case DefaultCodeFormatterOptions.TAB:
			boolean useTabsForLeadingIndents = useTabsOnlyForLeadingIndents;
			int numberOfLeadingIndents = numberOfIndentations;
			int indentationsAsTab = 0;
			if (useTabsForLeadingIndents) {
				while (column <= indentationLevel) {
					if (indentationsAsTab < numberOfLeadingIndents) {
						buffer.append('\t');
						indentationsAsTab++;
						int complement = tabLength - ((column - 1) % tabLength); // amount of space
						column += complement;
						needSpace = false;
					} else {
						buffer.append(' ');
						column++;
						needSpace = false;
					}
				}
			} else {
				while (column <= indentationLevel - indentationLevel % tabLength) {
					buffer.append('\t');
					int complement = tabLength - ((column - 1) % tabLength); // amount of space
					column += complement;
					needSpace = false;
				}
				while (column <= indentationLevel) {
					buffer.append(' ');
					column++;
					needSpace = false;
				}
			}
			break;
		case DefaultCodeFormatterOptions.SPACE:
			while (column <= indentationLevel) {
				buffer.append(' ');
				column++;
				needSpace = false;
			}
			break;
		case DefaultCodeFormatterOptions.MIXED:
			useTabsForLeadingIndents = useTabsOnlyForLeadingIndents;
			numberOfLeadingIndents = numberOfIndentations;
			indentationsAsTab = 0;
			if (useTabsForLeadingIndents) {
				final int columnForLeadingIndents = numberOfLeadingIndents * indentationSize;
				while (column <= indentationLevel) {
					if (column <= columnForLeadingIndents) {
						if ((column - 1 + tabLength) <= indentationLevel) {
							buffer.append('\t');
							column += tabLength;
						} else if ((column - 1 + indentationSize) <= indentationLevel) {
							// print one indentation
							for (int i = 0, max = indentationSize; i < max; i++) {
								buffer.append(' ');
								column++;
							}
						} else {
							buffer.append(' ');
							column++;
						}
					} else {
						for (int i = column, max = indentationLevel; i <= max; i++) {
							buffer.append(' ');
							column++;
						}
					}
					needSpace = false;
				}
			} else {
				while (column <= indentationLevel) {
					if ((column - 1 + tabLength) <= indentationLevel) {
						buffer.append('\t');
						column += tabLength;
					} else if ((column - 1 + indentationSize) <= indentationLevel) {
						// print one indentation
						for (int i = 0, max = indentationSize; i < max; i++) {
							buffer.append(' ');
							column++;
						}
					} else {
						buffer.append(' ');
						column++;
					}
					needSpace = false;
				}
			}
			break;
		}
	}

	public void enterNode() {
		lastLineComment.contiguous = false;
	}

	public void startNewLine() {
		if (column > 1) {
			printNewLine();
		}
	}

	public void printNewLine() {
		printNewLine(scanner.getCurrentPosition());
	}

	public void printNewLine(int insertPosition) {
		if (shouldSkip(scanner.getCurrentPosition())) {
			return;
		}
		if (lastNumberOfNewLines >= 1) {
			// Ensure that the scribe is at the beginning of a new line
			// only if no specific indentation has been previously set.
			if (!preserveLineBreakIndentation) {
				column = 1;
			}
			preserveLineBreakIndentation = false;
			return;
		}
		addInsertEdit(insertPosition, lineSeparator);
		line++;
		lastNumberOfNewLines = 1;
		column = 1;
		needSpace = false;
		pendingSpace = false;
		preserveLineBreakIndentation = false;
		lastLineComment.contiguous = false;
	}

	public void printNextToken(int expectedTokenType) {
		printNextToken(expectedTokenType, false);
	}

	public void printNextToken(int expectedTokenType, boolean considerSpaceIfAny) {
		// Set brace flag, it's useful for the scribe while preserving line breaks
		switch (expectedTokenType) {
		case Token.tRBRACE:
		case Token.tLBRACE:
			formatBrace = true;
			lastLineComment.contiguous = false;
		}
		try {
			printComment();
			if (shouldSkip(scanner.getCurrentPosition())) {
				return;
			}
			currentToken = scanner.nextToken();
			if (currentToken == null || expectedTokenType != currentToken.type) {
				if (pendingSpace) {
					addInsertEdit(scanner.getCurrentTokenStartPosition(), SPACE);
				}
				pendingSpace = false;
				needSpace = true;
				throw new AbortFormatting("[" + (line + 1) + "/" + column + "] Unexpected token type, expecting:" + //$NON-NLS-1$//$NON-NLS-2$//$NON-NLS-3$
						expectedTokenType + ", actual:" + currentToken);//$NON-NLS-1$
			}
			print(currentToken.getLength(), considerSpaceIfAny);
		} finally {
			formatBrace = false;
		}
	}

	public void printNextToken(int[] expectedTokenTypes) {
		printNextToken(expectedTokenTypes, false);
	}

	public void printNextToken(int[] expectedTokenTypes, boolean considerSpaceIfAny) {
		printComment();
		if (shouldSkip(scanner.getCurrentPosition())) {
			return;
		}
		currentToken = scanner.nextToken();
		if (Arrays.binarySearch(expectedTokenTypes, currentToken.type) < 0) {
			if (pendingSpace) {
				addInsertEdit(scanner.getCurrentTokenStartPosition(), SPACE);
			}
			pendingSpace = false;
			needSpace = true;
			StringBuilder expectations = new StringBuilder(5);
			for (int i = 0; i < expectedTokenTypes.length; i++) {
				if (i > 0) {
					expectations.append(',');
				}
				expectations.append(expectedTokenTypes[i]);
			}
			throw new AbortFormatting("[" + (line + 1) + "/" + column + "] unexpected token type, expecting:[" //$NON-NLS-1$//$NON-NLS-2$//$NON-NLS-3$
					+ expectations.toString() + "], actual:" + currentToken); //$NON-NLS-1$
		}
		print(currentToken.getLength(), considerSpaceIfAny);
	}

	private void printRuler(StringBuilder stringBuffer) {
		for (int i = 0; i < pageWidth; i++) {
			if ((i % tabLength) == 0) {
				stringBuffer.append('+');
			} else {
				stringBuffer.append('-');
			}
		}
		stringBuffer.append(lineSeparator);

		for (int i = 0; i < (pageWidth / tabLength); i++) {
			stringBuffer.append(i);
			stringBuffer.append('\t');
		}
	}

	public void printSpaces(int numSpaces) {
		if (numSpaces > 0) {
			int currentPosition = scanner.getCurrentPosition();
			StringBuilder spaces = new StringBuilder(numSpaces);
			for (int i = 0; i < numSpaces; i++) {
				spaces.append(' ');
			}
			addInsertEdit(currentPosition, spaces);
			pendingSpace = false;
			needSpace = false;
		}
	}

	public void printTrailingComment() {
		printComment(BASIC_TRAILING_COMMENT);
	}

	void redoAlignment(AlignmentException e) {
		if (e.relativeDepth > 0) { // if exception targets a distinct context
			e.relativeDepth--; // record fact that current context got traversed
			currentAlignment = currentAlignment.enclosing; // pop currentLocation
			throw e; // rethrow
		}
		// Reset scribe/scanner to restart at this given location
		resetAt(currentAlignment.location);
		scanner.resetTo(currentAlignment.location.inputOffset, scanner.eofPosition);
		// Clean alignment chunkKind so it will think it is a new chunk again
		currentAlignment.chunkKind = 0;
		currentAlignmentException = null;
	}

	void redoMemberAlignment(AlignmentException e) {
		// Reset scribe/scanner to restart at this given location
		resetAt(memberAlignment.location);
		scanner.resetTo(memberAlignment.location.inputOffset, scanner.eofPosition);
		// Clean alignment chunkKind so it will think it is a new chunk again
		memberAlignment.chunkKind = 0;
		currentAlignmentException = null;
	}

	public void reset() {
		checkLineWrapping = true;
		line = 0;
		column = 1;
		editsIndex = 0;
	}

	private void resetAt(Location location) {
		line = location.outputLine;
		column = location.outputColumn;
		indentationLevel = location.outputIndentationLevel;
		needSpace = location.needSpace;
		pendingSpace = location.pendingSpace;
		numberOfIndentations = location.numberOfIndentations;
		lastNumberOfNewLines = location.lastNumberOfNewLines;
		editsIndex = location.editsIndex;
		if (editsIndex > 0) {
			edits[editsIndex - 1] = location.textEdit;
		}
		setTailFormatter(location.tailFormatter);
	}

	private void resize() {
		System.arraycopy(edits, 0, (edits = new OptimizedReplaceEdit[editsIndex * 2]), 0, editsIndex);
	}

	public void space() {
		if (!needSpace)
			return;
		if (shouldSkip(scanner.getCurrentPosition())) {
			return;
		}
		lastNumberOfNewLines = 0;
		pendingSpace = true;
		column++;
		needSpace = false;
	}

	public void undoSpace() {
		if (pendingSpace) {
			pendingSpace = false;
			needSpace = true;
			column--;
		}
	}

	@Override
	public String toString() {
		StringBuilder buffer = new StringBuilder();
		buffer.append("(page width = ").append(pageWidth).append(") - (tabChar = "); //$NON-NLS-1$//$NON-NLS-2$
		switch (tabChar) {
		case DefaultCodeFormatterOptions.TAB:
			buffer.append("TAB"); //$NON-NLS-1$
			break;
		case DefaultCodeFormatterOptions.SPACE:
			buffer.append("SPACE"); //$NON-NLS-1$
			break;
		default:
			buffer.append("MIXED"); //$NON-NLS-1$
		}
		buffer.append(") - (tabSize = ").append(tabLength).append(')') //$NON-NLS-1$/
				.append(lineSeparator).append("(line = ").append(line).append(") - (column = ").append(column) //$NON-NLS-1$//$NON-NLS-2$
				.append(") - (identationLevel = ").append(indentationLevel).append(')') //$NON-NLS-1$
				.append(lineSeparator).append("(needSpace = ").append(needSpace).append(") - (lastNumberOfNewLines = ") //$NON-NLS-1$//$NON-NLS-2$
				.append(lastNumberOfNewLines).append(") - (checkLineWrapping = ").append(checkLineWrapping).append(')') //$NON-NLS-1$
				.append(lineSeparator)
				.append("==================================================================================") //$NON-NLS-1$
				.append(lineSeparator);
		printRuler(buffer);
		return buffer.toString();
	}

	public void unIndent() {
		if (shouldSkip(scanner.getCurrentPosition())) {
			fSkippedIndentations--;
			return;
		}
		indentationLevel -= indentationSize;
		numberOfIndentations--;
	}

	/**
	 */
	public boolean printModifiers() {
		int currentTokenStartPosition = scanner.getCurrentPosition();
		if (shouldSkip(currentTokenStartPosition)) {
			return false;
		}
		boolean isFirstModifier = true;
		boolean hasComment = false;
		while ((currentToken = scanner.nextToken()) != null) {
			switch (currentToken.type) {
			case Token.t_typedef:
			case Token.t_extern:
			case Token.t_static:
			case Token.t_auto:
			case Token.t_register:
			case Token.t_const:
			case Token.t_signed:
			case Token.t_unsigned:
			case Token.t_volatile:
			case Token.t_virtual:
			case Token.t_mutable:
			case Token.t_explicit:
			case Token.t_friend:
			case Token.t_inline:
			case Token.t_restrict:
				print(currentToken.getLength(), !isFirstModifier);
				isFirstModifier = false;
				currentTokenStartPosition = scanner.getCurrentPosition();
				break;
			case Token.tBLOCKCOMMENT:
				printBlockComment(false);
				currentTokenStartPosition = scanner.getCurrentPosition();
				hasComment = true;
				break;
			case Token.tLINECOMMENT:
				printLineComment();
				currentTokenStartPosition = scanner.getCurrentPosition();
				break;
			case Token.tWHITESPACE:
				addDeleteEdit(scanner.getCurrentTokenStartPosition(), scanner.getCurrentTokenEndPosition());
				int count = 0;
				char[] whiteSpaces = scanner.getCurrentTokenSource();
				for (int i = 0, max = whiteSpaces.length; i < max; i++) {
					switch (whiteSpaces[i]) {
					case '\r':
						if ((i + 1) < max) {
							if (whiteSpaces[i + 1] == '\n') {
								i++;
							}
						}
						count++;
						break;
					case '\n':
						count++;
					}
				}
				if (count >= 1 && hasComment) {
					printNewLine();
				}
				currentTokenStartPosition = scanner.getCurrentPosition();
				hasComment = false;
				break;
			case Token.tPREPROCESSOR:
			case Token.tPREPROCESSOR_DEFINE:
			case Token.tPREPROCESSOR_INCLUDE:
				if (column != 1)
					printNewLine(scanner.getCurrentTokenStartPosition());
				printPreprocessorDirective();
				printNewLine();
				currentTokenStartPosition = scanner.getCurrentPosition();
				hasComment = false;
				break;
			default:
				if (currentToken.getType() == Token.tIDENTIFIER) {
					if (currentToken.getText().startsWith("__")) { //$NON-NLS-1$
						// assume this is a declspec modifier
						print(currentToken.getLength(), !isFirstModifier);
						isFirstModifier = false;
						currentTokenStartPosition = scanner.getCurrentPosition();
						if ((currentToken = scanner.nextToken()) != null) {
							if (currentToken.getType() == Token.tLPAREN) {
								if (skipToToken(Token.tRPAREN)) {
									currentToken = scanner.nextToken();
									currentToken = scanner.nextToken();
									currentTokenStartPosition = scanner.getCurrentPosition();
								}
							}
						}
						break;
					}
				}
				// Step back one token
				scanner.resetTo(currentTokenStartPosition, scannerEndPosition);
				return !isFirstModifier;
			}
		}
		return !isFirstModifier;
	}

	/**
	 * Skips to the next occurrence of the given token type.
	 * If successful, the next token will be the expected token,
	 * otherwise the scanner position is left unchanged.
	 *
	 * @param expectedTokenType
	 * @return  <code>true</code> if a matching token was skipped to
	 */
	public boolean skipToToken(int expectedTokenType) {
		int skipStart = scanner.getCurrentPosition();
		if (shouldSkip(skipStart)) {
			return true;
		}
		int tokenStart = findToken(expectedTokenType);
		if (tokenStart < 0) {
			return false;
		}
		printRaw(skipStart, tokenStart - skipStart);
		currentToken = scanner.nextToken();
		scanner.resetTo(tokenStart, scannerEndPosition);
		return true;
	}

	/**
	 * Searches for the next occurrence of the given token type.
	 * If successful, returns the offset of the found token, otherwise -1.
	 * The scanner position is left unchanged.
	 *
	 * @param tokenType type of the token to look for
	 * @return the position of the matching token, if found, otherwise -1.
	 */
	public int findToken(int tokenType) {
		return findToken(tokenType, scannerEndPosition - 1);
	}

	/**
	 * Searches for the next occurrence of the given token type.
	 * If successful, returns the offset of the found token, otherwise -1.
	 * The scanner position is left unchanged.
	 *
	 * @param tokenType type of the token to look for
	 * @param endPosition end position limiting the search
	 * @return the position of the matching token, if found, otherwise -1.
	 */
	public int findToken(int tokenType, int endPosition) {
		int startPosition = scanner.getCurrentPosition();
		if (startPosition >= endPosition) {
			return -1;
		}
		try {
			int braceLevel = 0;
			int parenLevel = 0;
			switch (tokenType) {
			case Token.tRBRACE:
				++braceLevel;
				break;
			case Token.tRPAREN:
				++parenLevel;
				break;
			}
			Token token;
			while ((token = scanner.nextToken()) != null) {
				if (scanner.getCurrentTokenEndPosition() > endPosition)
					return -1;

				switch (token.type) {
				case Token.tLBRACE:
					if (tokenType != Token.tLBRACE) {
						++braceLevel;
					}
					break;
				case Token.tRBRACE:
					--braceLevel;
					break;
				case Token.tLPAREN:
					if (tokenType != Token.tLPAREN) {
						++parenLevel;
					}
					break;
				case Token.tRPAREN:
					--parenLevel;
					break;
				case Token.tWHITESPACE:
				case Token.tLINECOMMENT:
				case Token.tBLOCKCOMMENT:
				case Token.tPREPROCESSOR:
				case Token.tPREPROCESSOR_DEFINE:
				case Token.tPREPROCESSOR_INCLUDE:
					continue;
				}
				if (braceLevel <= 0 && parenLevel <= 0) {
					if (token.type == tokenType) {
						return scanner.getCurrentTokenStartPosition();
					}
				}
				if (braceLevel < 0 || parenLevel < 0) {
					break;
				}
			}
		} finally {
			scanner.resetTo(startPosition, scannerEndPosition);
		}
		return -1;
	}

	/**
	 * Searches for the next occurrence of the given token type.
	 * If successful, returns the offset of the found token, otherwise -1.
	 * The scanner position is left unchanged.
	 *
	 * @param tokenType type of the token to look for
	 * @param startPosition position where to start the search
	 * @param endPosition end position limiting the search
	 * @return the position of the matching token, if found, otherwise -1.
	 */
	public int findToken(int tokenType, int startPosition, int endPosition) {
		int currentPosition = scanner.getCurrentPosition();
		try {
			scanner.resetTo(startPosition, scannerEndPosition);
			return findToken(tokenType, endPosition);
		} finally {
			scanner.resetTo(currentPosition, scannerEndPosition);
		}
	}

	public boolean printCommentPreservingNewLines() {
		final boolean savedPreserveNL = preserveNewLines;
		preserveNewLines = true;
		try {
			return printComment();
		} finally {
			preserveNewLines = savedPreserveNL;
		}
	}

	boolean shouldSkip(int offset) {
		return offset >= fSkipStartOffset && offset < fSkipEndOffset;
	}

	void skipRange(int offset, int endOffset) {
		if (offset == fSkipStartOffset) {
			return;
		}
		final int currentPosition = scanner.getCurrentPosition();
		if (offset > currentPosition) {
			fSkipStartOffset = Integer.MAX_VALUE;
			printRaw(currentPosition, offset - currentPosition);
		}
		fSkipStartOffset = offset;
		fSkipEndOffset = endOffset;
	}

	boolean skipRange() {
		return fSkipEndOffset > 0;
	}

	void restartAtOffset(int offset) {
		final int currentPosition = scanner.getCurrentPosition();
		if (fSkipEndOffset > 0) {
			fSkipStartOffset = Integer.MAX_VALUE;
			fSkipEndOffset = 0;
			while (fSkippedIndentations < 0) {
				unIndent();
				fSkippedIndentations++;
			}
			if (offset > currentPosition) {
				printRaw(currentPosition, offset - currentPosition);
				scanner.resetTo(offset, scannerEndPosition);
			}
			while (fSkippedIndentations > 0) {
				indent();
				fSkippedIndentations--;
			}
		} else if (offset > currentPosition) {
			boolean hasSpace = printComment();
			final int nextPosition = scanner.getCurrentPosition();
			if (offset > nextPosition) {
				if (hasSpace) {
					space();
				}
				printRaw(nextPosition, offset - nextPosition);
				scanner.resetTo(offset, scannerEndPosition);
			}
		}
	}

	/*
	 * Returns the tail formatter associated with the current alignment or, if there is no current
	 * alignment, with the scribe itself.
	 * @see #tailFormatter
	 */
	public Runnable getTailFormatter() {
		if (currentAlignment != null) {
			return currentAlignment.tailFormatter;
		} else {
			return this.tailFormatter;
		}
	}

	/*
	 * Returns the tail formatter associated with the current alignment or, if there is no current
	 * alignment, with the scribe itself. The tail formatter associated with the alignment or
	 * the scribe is set to {@code null}.
	 * @see #tailFormatter
	 */
	public Runnable takeTailFormatter() {
		Runnable formatter;
		if (currentAlignment != null) {
			formatter = currentAlignment.tailFormatter;
			currentAlignment.tailFormatter = null;
		} else {
			formatter = this.tailFormatter;
			this.tailFormatter = null;
		}
		return formatter;
	}

	/*
	 * Sets the tail formatter associated with the current alignment or, if there is no current
	 * alignment, with the scribe itself.
	 * @see #tailFormatter
	 */
	public void setTailFormatter(Runnable tailFormatter) {
		if (currentAlignment != null) {
			currentAlignment.tailFormatter = tailFormatter;
		} else {
			this.tailFormatter = tailFormatter;
		}
	}

	/*
	 * Runs the tail formatter associated with the current alignment or, if there is no current
	 * alignment, with the scribe itself.
	 * @see #tailFormatter
	 */
	public void runTailFormatter() {
		Runnable formatter = getTailFormatter();
		if (formatter != null)
			formatter.run();
	}
}