mcimadamore@1113: /*
jjg@1339:  * Copyright (c) 2011, 2012, Oracle and/or its affiliates. All rights reserved.
mcimadamore@1113:  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
mcimadamore@1113:  *
mcimadamore@1113:  * This code is free software; you can redistribute it and/or modify it
mcimadamore@1113:  * under the terms of the GNU General Public License version 2 only, as
mcimadamore@1113:  * published by the Free Software Foundation.  Oracle designates this
mcimadamore@1113:  * particular file as subject to the "Classpath" exception as provided
mcimadamore@1113:  * by Oracle in the LICENSE file that accompanied this code.
mcimadamore@1113:  *
mcimadamore@1113:  * This code is distributed in the hope that it will be useful, but WITHOUT
mcimadamore@1113:  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
mcimadamore@1113:  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
mcimadamore@1113:  * version 2 for more details (a copy is included in the LICENSE file that
mcimadamore@1113:  * accompanied this code).
mcimadamore@1113:  *
mcimadamore@1113:  * You should have received a copy of the GNU General Public License version
mcimadamore@1113:  * 2 along with this work; if not, write to the Free Software Foundation,
mcimadamore@1113:  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
mcimadamore@1113:  *
mcimadamore@1113:  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
mcimadamore@1113:  * or visit www.oracle.com if you need additional information or have any
mcimadamore@1113:  * questions.
mcimadamore@1113:  */
mcimadamore@1113: 
mcimadamore@1113: package com.sun.tools.javac.parser;
mcimadamore@1113: 
jjg@1339: import java.nio.CharBuffer;
jjg@1339: import java.util.Arrays;
jjg@1339: 
mcimadamore@1113: import com.sun.tools.javac.file.JavacFileManager;
jjg@1339: import com.sun.tools.javac.util.ArrayUtils;
mcimadamore@1125: import com.sun.tools.javac.util.Log;
mcimadamore@1125: import com.sun.tools.javac.util.Name;
mcimadamore@1125: import com.sun.tools.javac.util.Names;
mcimadamore@1125: 
mcimadamore@1113: import static com.sun.tools.javac.util.LayoutCharacters.*;
mcimadamore@1113: 
mcimadamore@1113: /** The char reader used by the javac lexer/tokenizer. Returns the sequence of
mcimadamore@1113:  * characters contained in the input stream, handling unicode escape accordingly.
mcimadamore@1113:  * Additionally, it provide features for saving chars into a buffer and to retrieve
mcimadamore@1113:  * them at a later stage.
mcimadamore@1113:  *
mcimadamore@1113:  *  <p><b>This is NOT part of any supported API.
mcimadamore@1113:  *  If you write code that depends on this, you do so at your own risk.
mcimadamore@1113:  *  This code and its internal interfaces are subject to change or
mcimadamore@1113:  *  deletion without notice.</b>
mcimadamore@1113:  */
mcimadamore@1113: public class UnicodeReader {
mcimadamore@1113: 
mcimadamore@1113:     /** The input buffer, index of next character to be read,
mcimadamore@1113:      *  index of one past last character in buffer.
mcimadamore@1113:      */
mcimadamore@1113:     protected char[] buf;
mcimadamore@1113:     protected int bp;
mcimadamore@1113:     protected final int buflen;
mcimadamore@1113: 
mcimadamore@1113:     /** The current character.
mcimadamore@1113:      */
mcimadamore@1113:     protected char ch;
mcimadamore@1113: 
mcimadamore@1113:     /** The buffer index of the last converted unicode character
mcimadamore@1113:      */
mcimadamore@1113:     protected int unicodeConversionBp = -1;
mcimadamore@1113: 
mcimadamore@1113:     protected Log log;
mcimadamore@1125:     protected Names names;
mcimadamore@1125: 
mcimadamore@1125:     /** A character buffer for saved chars.
mcimadamore@1125:      */
mcimadamore@1125:     protected char[] sbuf = new char[128];
mcimadamore@1125:     protected int sp;
mcimadamore@1113: 
mcimadamore@1113:     /**
mcimadamore@1113:      * Create a scanner from the input array.  This method might
mcimadamore@1113:      * modify the array.  To avoid copying the input array, ensure
mcimadamore@1113:      * that {@code inputLength < input.length} or
mcimadamore@1113:      * {@code input[input.length -1]} is a white space character.
mcimadamore@1113:      *
jjg@1358:      * @param sf the factory which created this Scanner
jjg@1358:      * @param buffer the input, might be modified
mcimadamore@1113:      * Must be positive and less than or equal to input.length.
mcimadamore@1113:      */
mcimadamore@1113:     protected UnicodeReader(ScannerFactory sf, CharBuffer buffer) {
mcimadamore@1113:         this(sf, JavacFileManager.toArray(buffer), buffer.limit());
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     protected UnicodeReader(ScannerFactory sf, char[] input, int inputLength) {
mcimadamore@1113:         log = sf.log;
mcimadamore@1125:         names = sf.names;
mcimadamore@1113:         if (inputLength == input.length) {
mcimadamore@1113:             if (input.length > 0 && Character.isWhitespace(input[input.length - 1])) {
mcimadamore@1113:                 inputLength--;
mcimadamore@1113:             } else {
jjg@1339:                 input = Arrays.copyOf(input, inputLength + 1);
mcimadamore@1113:             }
mcimadamore@1113:         }
mcimadamore@1113:         buf = input;
mcimadamore@1113:         buflen = inputLength;
mcimadamore@1113:         buf[buflen] = EOI;
mcimadamore@1113:         bp = -1;
mcimadamore@1113:         scanChar();
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     /** Read next character.
mcimadamore@1113:      */
mcimadamore@1113:     protected void scanChar() {
mcimadamore@1113:         if (bp < buflen) {
mcimadamore@1113:             ch = buf[++bp];
mcimadamore@1113:             if (ch == '\\') {
mcimadamore@1113:                 convertUnicode();
mcimadamore@1113:             }
mcimadamore@1113:         }
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1125:     /** Read next character in comment, skipping over double '\' characters.
mcimadamore@1125:      */
mcimadamore@1125:     protected void scanCommentChar() {
mcimadamore@1125:         scanChar();
mcimadamore@1125:         if (ch == '\\') {
mcimadamore@1125:             if (peekChar() == '\\' && !isUnicode()) {
mcimadamore@1125:                 skipChar();
mcimadamore@1125:             } else {
mcimadamore@1125:                 convertUnicode();
mcimadamore@1125:             }
mcimadamore@1125:         }
mcimadamore@1125:     }
mcimadamore@1125: 
mcimadamore@1125:     /** Append a character to sbuf.
mcimadamore@1125:      */
mcimadamore@1125:     protected void putChar(char ch, boolean scan) {
jjg@1339:         sbuf = ArrayUtils.ensureCapacity(sbuf, sp);
mcimadamore@1125:         sbuf[sp++] = ch;
mcimadamore@1125:         if (scan)
mcimadamore@1125:             scanChar();
mcimadamore@1125:     }
mcimadamore@1125: 
mcimadamore@1125:     protected void putChar(char ch) {
mcimadamore@1125:         putChar(ch, false);
mcimadamore@1125:     }
mcimadamore@1125: 
mcimadamore@1125:     protected void putChar(boolean scan) {
mcimadamore@1125:         putChar(ch, scan);
mcimadamore@1125:     }
mcimadamore@1125: 
mcimadamore@1125:     Name name() {
mcimadamore@1125:         return names.fromChars(sbuf, 0, sp);
mcimadamore@1125:     }
mcimadamore@1125: 
mcimadamore@1125:     String chars() {
mcimadamore@1125:         return new String(sbuf, 0, sp);
mcimadamore@1125:     }
mcimadamore@1125: 
mcimadamore@1113:     /** Convert unicode escape; bp points to initial '\' character
mcimadamore@1113:      *  (Spec 3.3).
mcimadamore@1113:      */
mcimadamore@1113:     protected void convertUnicode() {
mcimadamore@1113:         if (ch == '\\' && unicodeConversionBp != bp) {
mcimadamore@1113:             bp++; ch = buf[bp];
mcimadamore@1113:             if (ch == 'u') {
mcimadamore@1113:                 do {
mcimadamore@1113:                     bp++; ch = buf[bp];
mcimadamore@1113:                 } while (ch == 'u');
mcimadamore@1113:                 int limit = bp + 3;
mcimadamore@1113:                 if (limit < buflen) {
mcimadamore@1113:                     int d = digit(bp, 16);
mcimadamore@1113:                     int code = d;
mcimadamore@1113:                     while (bp < limit && d >= 0) {
mcimadamore@1113:                         bp++; ch = buf[bp];
mcimadamore@1113:                         d = digit(bp, 16);
mcimadamore@1113:                         code = (code << 4) + d;
mcimadamore@1113:                     }
mcimadamore@1113:                     if (d >= 0) {
mcimadamore@1113:                         ch = (char)code;
mcimadamore@1113:                         unicodeConversionBp = bp;
mcimadamore@1113:                         return;
mcimadamore@1113:                     }
mcimadamore@1113:                 }
mcimadamore@1113:                 log.error(bp, "illegal.unicode.esc");
mcimadamore@1113:             } else {
mcimadamore@1113:                 bp--;
mcimadamore@1113:                 ch = '\\';
mcimadamore@1113:             }
mcimadamore@1113:         }
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     /** Are surrogates supported?
mcimadamore@1113:      */
mcimadamore@1113:     final static boolean surrogatesSupported = surrogatesSupported();
mcimadamore@1113:     private static boolean surrogatesSupported() {
mcimadamore@1113:         try {
mcimadamore@1113:             Character.isHighSurrogate('a');
mcimadamore@1113:             return true;
mcimadamore@1113:         } catch (NoSuchMethodError ex) {
mcimadamore@1113:             return false;
mcimadamore@1113:         }
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     /** Scan surrogate pairs.  If 'ch' is a high surrogate and
mcimadamore@1113:      *  the next character is a low surrogate, then put the low
mcimadamore@1113:      *  surrogate in 'ch', and return the high surrogate.
mcimadamore@1113:      *  otherwise, just return 0.
mcimadamore@1113:      */
mcimadamore@1113:     protected char scanSurrogates() {
mcimadamore@1113:         if (surrogatesSupported && Character.isHighSurrogate(ch)) {
mcimadamore@1113:             char high = ch;
mcimadamore@1113: 
mcimadamore@1113:             scanChar();
mcimadamore@1113: 
mcimadamore@1113:             if (Character.isLowSurrogate(ch)) {
mcimadamore@1113:                 return high;
mcimadamore@1113:             }
mcimadamore@1113: 
mcimadamore@1113:             ch = high;
mcimadamore@1113:         }
mcimadamore@1113: 
mcimadamore@1113:         return 0;
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     /** Convert an ASCII digit from its base (8, 10, or 16)
mcimadamore@1113:      *  to its value.
mcimadamore@1113:      */
mcimadamore@1113:     protected int digit(int pos, int base) {
mcimadamore@1113:         char c = ch;
mcimadamore@1113:         int result = Character.digit(c, base);
mcimadamore@1113:         if (result >= 0 && c > 0x7f) {
mcimadamore@1113:             log.error(pos + 1, "illegal.nonascii.digit");
mcimadamore@1113:             ch = "0123456789abcdef".charAt(result);
mcimadamore@1113:         }
mcimadamore@1113:         return result;
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     protected boolean isUnicode() {
mcimadamore@1113:         return unicodeConversionBp == bp;
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     protected void skipChar() {
mcimadamore@1113:         bp++;
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     protected char peekChar() {
mcimadamore@1113:         return buf[bp + 1];
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     /**
mcimadamore@1113:      * Returns a copy of the input buffer, up to its inputLength.
mcimadamore@1113:      * Unicode escape sequences are not translated.
mcimadamore@1113:      */
mcimadamore@1113:     public char[] getRawCharacters() {
mcimadamore@1113:         char[] chars = new char[buflen];
mcimadamore@1113:         System.arraycopy(buf, 0, chars, 0, buflen);
mcimadamore@1113:         return chars;
mcimadamore@1113:     }
mcimadamore@1113: 
mcimadamore@1113:     /**
mcimadamore@1113:      * Returns a copy of a character array subset of the input buffer.
jjg@1358:      * The returned array begins at the {@code beginIndex} and
jjg@1358:      * extends to the character at index {@code endIndex - 1}.
jjg@1358:      * Thus the length of the substring is {@code endIndex-beginIndex}.
mcimadamore@1113:      * This behavior is like
jjg@1358:      * {@code String.substring(beginIndex, endIndex)}.
mcimadamore@1113:      * Unicode escape sequences are not translated.
mcimadamore@1113:      *
mcimadamore@1113:      * @param beginIndex the beginning index, inclusive.
mcimadamore@1113:      * @param endIndex the ending index, exclusive.
jjg@1358:      * @throws ArrayIndexOutOfBoundsException if either offset is outside of the
mcimadamore@1113:      *         array bounds
mcimadamore@1113:      */
mcimadamore@1113:     public char[] getRawCharacters(int beginIndex, int endIndex) {
mcimadamore@1113:         int length = endIndex - beginIndex;
mcimadamore@1113:         char[] chars = new char[length];
mcimadamore@1113:         System.arraycopy(buf, beginIndex, chars, 0, length);
mcimadamore@1113:         return chars;
mcimadamore@1113:     }
mcimadamore@1113: }