aoqi@0: /*
aoqi@0:  * Copyright (c) 2009, Oracle and/or its affiliates. All rights reserved.
aoqi@0:  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
aoqi@0:  *
aoqi@0:  * This code is free software; you can redistribute it and/or modify it
aoqi@0:  * under the terms of the GNU General Public License version 2 only, as
aoqi@0:  * published by the Free Software Foundation.  Oracle designates this
aoqi@0:  * particular file as subject to the "Classpath" exception as provided
aoqi@0:  * by Oracle in the LICENSE file that accompanied this code.
aoqi@0:  *
aoqi@0:  * This code is distributed in the hope that it will be useful, but WITHOUT
aoqi@0:  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
aoqi@0:  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
aoqi@0:  * version 2 for more details (a copy is included in the LICENSE file that
aoqi@0:  * accompanied this code).
aoqi@0:  *
aoqi@0:  * You should have received a copy of the GNU General Public License version
aoqi@0:  * 2 along with this work; if not, write to the Free Software Foundation,
aoqi@0:  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
aoqi@0:  *
aoqi@0:  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
aoqi@0:  * or visit www.oracle.com if you need additional information or have any
aoqi@0:  * questions.
aoqi@0:  */
aoqi@0: 
aoqi@0: package com.sun.xml.internal.dtdparser;
aoqi@0: 
aoqi@0: 
aoqi@0: /**
aoqi@0:  * Methods in this class are used to determine whether characters may
aoqi@0:  * appear in certain roles in XML documents.  Such methods are used
aoqi@0:  * both to parse and to create such documents.
aoqi@0:  *
aoqi@0:  * @author David Brownell
aoqi@0:  * @version 1.1, 00/08/05
aoqi@0:  */
aoqi@0: public class XmlChars {
aoqi@0:     // can't construct instances
aoqi@0:     private XmlChars() {
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Returns true if the argument, a UCS-4 character code, is valid in
aoqi@0:      * XML documents.  Unicode characters fit into the low sixteen
aoqi@0:      * bits of a UCS-4 character, and pairs of Unicode <em>surrogate
aoqi@0:      * characters</em> can be combined to encode UCS-4 characters in
aoqi@0:      * documents containing only Unicode.  (The <code>char</code> datatype
aoqi@0:      * in the Java Programming Language represents Unicode characters,
aoqi@0:      * including unpaired surrogates.)
aoqi@0:      * <p/>
aoqi@0:      * <P> In XML, UCS-4 characters can also be encoded by the use of
aoqi@0:      * <em>character references</em> such as <b>&amp;#x12345678;</b>, which
aoqi@0:      * happens to refer to a character that is disallowed in XML documents.
aoqi@0:      * UCS-4 characters allowed in XML documents can be expressed with
aoqi@0:      * one or two Unicode characters.
aoqi@0:      *
aoqi@0:      * @param ucs4char The 32-bit UCS-4 character being tested.
aoqi@0:      */
aoqi@0:     static public boolean isChar(int ucs4char) {
aoqi@0:         // [2] Char ::= #x0009 | #x000A | #x000D
aoqi@0:         //            | [#x0020-#xD7FF]
aoqi@0:         //    ... surrogates excluded!
aoqi@0:         //            | [#xE000-#xFFFD]
aoqi@0:         //             | [#x10000-#x10ffff]
aoqi@0:         return ((ucs4char >= 0x0020 && ucs4char <= 0xD7FF)
aoqi@0:                 || ucs4char == 0x000A || ucs4char == 0x0009
aoqi@0:                 || ucs4char == 0x000D
aoqi@0:                 || (ucs4char >= 0xE000 && ucs4char <= 0xFFFD)
aoqi@0:                 || (ucs4char >= 0x10000 && ucs4char <= 0x10ffff));
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Returns true if the character is allowed to be a non-initial
aoqi@0:      * character in names according to the XML recommendation.
aoqi@0:      *
aoqi@0:      * @see #isNCNameChar(char)
aoqi@0:      * @see #isLetter(char)
aoqi@0:      */
aoqi@0:     public static boolean isNameChar(char c) {
aoqi@0:         // [4] NameChar ::= Letter | Digit | '.' | '_' | ':'
aoqi@0:         //            | CombiningChar | Extender
aoqi@0: 
aoqi@0:         if (isLetter2(c))
aoqi@0:             return true;
aoqi@0:         else if (c == '>')
aoqi@0:             return false;
aoqi@0:         else if (c == '.' || c == '-' || c == '_' || c == ':'
aoqi@0:                 || isExtender(c))
aoqi@0:             return true;
aoqi@0:         else
aoqi@0:             return false;
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Returns true if the character is allowed to be a non-initial
aoqi@0:      * character in unscoped names according to the rules of the XML
aoqi@0:      * Namespaces proposed recommendation.  Except for precluding
aoqi@0:      * the colon (used to separate names from their scopes) these
aoqi@0:      * characters are just as allowed by the XML recommendation.
aoqi@0:      *
aoqi@0:      * @see #isNameChar(char)
aoqi@0:      * @see #isLetter(char)
aoqi@0:      */
aoqi@0:     public static boolean isNCNameChar(char c) {
aoqi@0:         // [NC 5] NCNameChar ::= Letter | Digit | '.' | '_'
aoqi@0:         //            | CombiningChar | Extender
aoqi@0:         return c != ':' && isNameChar(c);
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Returns true if the character is allowed where XML supports
aoqi@0:      * whitespace characters, false otherwise.
aoqi@0:      */
aoqi@0:     public static boolean isSpace(char c) {
aoqi@0:         return c == ' ' || c == '\t' || c == '\n' || c == '\r';
aoqi@0:     }
aoqi@0: 
aoqi@0: 
aoqi@0:     /*
aoqi@0:      * NOTE:  java.lang.Character.getType() values are:
aoqi@0:      *
aoqi@0:      * UNASSIGNED                    = 0,
aoqi@0:      *
aoqi@0:      * UPPERCASE_LETTER            = 1,    // Lu
aoqi@0:      * LOWERCASE_LETTER            = 2,    // Ll
aoqi@0:      * TITLECASE_LETTER            = 3,    // Lt
aoqi@0:      * MODIFIER_LETTER             = 4,    // Lm
aoqi@0:      * OTHER_LETTER                = 5,    // Lo
aoqi@0:      * NON_SPACING_MARK            = 6,    // Mn
aoqi@0:      * ENCLOSING_MARK              = 7,    // Me
aoqi@0:      * COMBINING_SPACING_MARK      = 8,    // Mc
aoqi@0:      * DECIMAL_DIGIT_NUMBER        = 9,    // Nd
aoqi@0:      * LETTER_NUMBER               = 10,   // Nl
aoqi@0:      * OTHER_NUMBER                = 11,   // No
aoqi@0:      * SPACE_SEPARATOR             = 12,   // Zs
aoqi@0:      * LINE_SEPARATOR              = 13,   // Zl
aoqi@0:      * PARAGRAPH_SEPARATOR         = 14,   // Zp
aoqi@0:      * CONTROL                     = 15,   // Cc
aoqi@0:      * FORMAT                      = 16,   // Cf
aoqi@0:      *                         // 17 reserved for proposed Ci category
aoqi@0:      * PRIVATE_USE                 = 18,   // Co
aoqi@0:      * SURROGATE                   = 19,   // Cs
aoqi@0:      * DASH_PUNCTUATION            = 20,   // Pd
aoqi@0:      * START_PUNCTUATION           = 21,   // Ps
aoqi@0:      * END_PUNCTUATION             = 22,   // Pe
aoqi@0:      * CONNECTOR_PUNCTUATION       = 23,   // Pc
aoqi@0:      * OTHER_PUNCTUATION           = 24,   // Po
aoqi@0:      * MATH_SYMBOL                 = 25,   // Sm
aoqi@0:      * CURRENCY_SYMBOL             = 26,   // Sc
aoqi@0:      * MODIFIER_SYMBOL             = 27,   // Sk
aoqi@0:      * OTHER_SYMBOL                = 28;   // So
aoqi@0:      */
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Returns true if the character is an XML "letter".  XML Names must
aoqi@0:      * start with Letters or a few other characters, but other characters
aoqi@0:      * in names must only satisfy the <em>isNameChar</em> predicate.
aoqi@0:      *
aoqi@0:      * @see #isNameChar(char)
aoqi@0:      * @see #isNCNameChar(char)
aoqi@0:      */
aoqi@0:     public static boolean isLetter(char c) {
aoqi@0:         // [84] Letter ::= BaseChar | Ideographic
aoqi@0:         // [85] BaseChar ::= ... too much to repeat
aoqi@0:         // [86] Ideographic ::= ... too much to repeat
aoqi@0: 
aoqi@0:         //
aoqi@0:         // Optimize the typical case.
aoqi@0:         //
aoqi@0:         if (c >= 'a' && c <= 'z')
aoqi@0:             return true;
aoqi@0:         if (c == '/')
aoqi@0:             return false;
aoqi@0:         if (c >= 'A' && c <= 'Z')
aoqi@0:             return true;
aoqi@0: 
aoqi@0:         //
aoqi@0:         // Since the tables are too ridiculous to use in code,
aoqi@0:         // we're using the footnotes here to drive this test.
aoqi@0:         //
aoqi@0:         switch (Character.getType(c)) {
aoqi@0:         // app. B footnote says these are 'name start'
aoqi@0:         // chars' ...
aoqi@0:         case Character.LOWERCASE_LETTER:        // Ll
aoqi@0:         case Character.UPPERCASE_LETTER:        // Lu
aoqi@0:         case Character.OTHER_LETTER:            // Lo
aoqi@0:         case Character.TITLECASE_LETTER:        // Lt
aoqi@0:         case Character.LETTER_NUMBER:            // Nl
aoqi@0: 
aoqi@0:             // OK, here we just have some exceptions to check...
aoqi@0:             return !isCompatibilityChar(c)
aoqi@0:                     // per "5.14 of Unicode", rule out some combiners
aoqi@0:                     && !(c >= 0x20dd && c <= 0x20e0);
aoqi@0: 
aoqi@0:         default:
aoqi@0:             // check for some exceptions:  these are "alphabetic"
aoqi@0:             return ((c >= 0x02bb && c <= 0x02c1)
aoqi@0:                     || c == 0x0559 || c == 0x06e5 || c == 0x06e6);
aoqi@0:         }
aoqi@0:     }
aoqi@0: 
aoqi@0:     //
aoqi@0:     // XML 1.0 discourages "compatibility" characters in names; these
aoqi@0:     // were defined to permit passing through some information stored in
aoqi@0:     // older non-Unicode character sets.  These always have alternative
aoqi@0:     // representations in Unicode, e.g. using combining chars.
aoqi@0:     //
aoqi@0:     private static boolean isCompatibilityChar(char c) {
aoqi@0:         // the numerous comparisions here seem unavoidable,
aoqi@0:         // but the switch can reduce the number which must
aoqi@0:         // actually be executed.
aoqi@0: 
aoqi@0:         switch ((c >> 8) & 0x0ff) {
aoqi@0:         case 0x00:
aoqi@0:             // ISO Latin/1 has a few compatibility characters
aoqi@0:             return c == 0x00aa || c == 0x00b5 || c == 0x00ba;
aoqi@0: 
aoqi@0:         case 0x01:
aoqi@0:             // as do Latin Extended A and (parts of) B
aoqi@0:             return (c >= 0x0132 && c <= 0x0133)
aoqi@0:                     || (c >= 0x013f && c <= 0x0140)
aoqi@0:                     || c == 0x0149
aoqi@0:                     || c == 0x017f
aoqi@0:                     || (c >= 0x01c4 && c <= 0x01cc)
aoqi@0:                     || (c >= 0x01f1 && c <= 0x01f3);
aoqi@0: 
aoqi@0:         case 0x02:
aoqi@0:             // some spacing modifiers
aoqi@0:             return (c >= 0x02b0 && c <= 0x02b8)
aoqi@0:                     || (c >= 0x02e0 && c <= 0x02e4);
aoqi@0: 
aoqi@0:         case 0x03:
aoqi@0:             return c == 0x037a;            // Greek
aoqi@0: 
aoqi@0:         case 0x05:
aoqi@0:             return c == 0x0587;            // Armenian
aoqi@0: 
aoqi@0:         case 0x0e:
aoqi@0:             return c >= 0x0edc && c <= 0x0edd;    // Laotian
aoqi@0: 
aoqi@0:         case 0x11:
aoqi@0:             // big chunks of Hangul Jamo are all "compatibility"
aoqi@0:             return c == 0x1101
aoqi@0:                     || c == 0x1104
aoqi@0:                     || c == 0x1108
aoqi@0:                     || c == 0x110a
aoqi@0:                     || c == 0x110d
aoqi@0:                     || (c >= 0x1113 && c <= 0x113b)
aoqi@0:                     || c == 0x113d
aoqi@0:                     || c == 0x113f
aoqi@0:                     || (c >= 0x1141 && c <= 0x114b)
aoqi@0:                     || c == 0x114d
aoqi@0:                     || c == 0x114f
aoqi@0:                     || (c >= 0x1151 && c <= 0x1153)
aoqi@0:                     || (c >= 0x1156 && c <= 0x1158)
aoqi@0:                     || c == 0x1162
aoqi@0:                     || c == 0x1164
aoqi@0:                     || c == 0x1166
aoqi@0:                     || c == 0x1168
aoqi@0:                     || (c >= 0x116a && c <= 0x116c)
aoqi@0:                     || (c >= 0x116f && c <= 0x1171)
aoqi@0:                     || c == 0x1174
aoqi@0:                     || (c >= 0x1176 && c <= 0x119d)
aoqi@0:                     || (c >= 0x119f && c <= 0x11a2)
aoqi@0:                     || (c >= 0x11a9 && c <= 0x11aa)
aoqi@0:                     || (c >= 0x11ac && c <= 0x11ad)
aoqi@0:                     || (c >= 0x11b0 && c <= 0x11b6)
aoqi@0:                     || c == 0x11b9
aoqi@0:                     || c == 0x11bb
aoqi@0:                     || (c >= 0x11c3 && c <= 0x11ea)
aoqi@0:                     || (c >= 0x11ec && c <= 0x11ef)
aoqi@0:                     || (c >= 0x11f1 && c <= 0x11f8)
aoqi@0:                     ;
aoqi@0: 
aoqi@0:         case 0x20:
aoqi@0:             return c == 0x207f;            // superscript
aoqi@0: 
aoqi@0:         case 0x21:
aoqi@0:             return
aoqi@0:                     // various letterlike symbols
aoqi@0:                     c == 0x2102
aoqi@0:                     || c == 0x2107
aoqi@0:                     || (c >= 0x210a && c <= 0x2113)
aoqi@0:                     || c == 0x2115
aoqi@0:                     || (c >= 0x2118 && c <= 0x211d)
aoqi@0:                     || c == 0x2124
aoqi@0:                     || c == 0x2128
aoqi@0:                     || (c >= 0x212c && c <= 0x212d)
aoqi@0:                     || (c >= 0x212f && c <= 0x2138)
aoqi@0: 
aoqi@0:                     // most Roman numerals (less 1K, 5K, 10K)
aoqi@0:                     || (c >= 0x2160 && c <= 0x217f)
aoqi@0:                     ;
aoqi@0: 
aoqi@0:         case 0x30:
aoqi@0:             // some Hiragana
aoqi@0:             return c >= 0x309b && c <= 0x309c;
aoqi@0: 
aoqi@0:         case 0x31:
aoqi@0:             // all Hangul Compatibility Jamo
aoqi@0:             return c >= 0x3131 && c <= 0x318e;
aoqi@0: 
aoqi@0:         case 0xf9:
aoqi@0:         case 0xfa:
aoqi@0:         case 0xfb:
aoqi@0:         case 0xfc:
aoqi@0:         case 0xfd:
aoqi@0:         case 0xfe:
aoqi@0:         case 0xff:
aoqi@0:             // the whole "compatibility" area is for that purpose!
aoqi@0:             return true;
aoqi@0: 
aoqi@0:         default:
aoqi@0:             // most of Unicode isn't flagged as being for compatibility
aoqi@0:             return false;
aoqi@0:         }
aoqi@0:     }
aoqi@0: 
aoqi@0:     // guts of isNameChar/isNCNameChar
aoqi@0:     private static boolean isLetter2(char c) {
aoqi@0:         // [84] Letter ::= BaseChar | Ideographic
aoqi@0:         // [85] BaseChar ::= ... too much to repeat
aoqi@0:         // [86] Ideographic ::= ... too much to repeat
aoqi@0:         // [87] CombiningChar ::= ... too much to repeat
aoqi@0: 
aoqi@0:         //
aoqi@0:         // Optimize the typical case.
aoqi@0:         //
aoqi@0:         if (c >= 'a' && c <= 'z')
aoqi@0:             return true;
aoqi@0:         if (c == '>')
aoqi@0:             return false;
aoqi@0:         if (c >= 'A' && c <= 'Z')
aoqi@0:             return true;
aoqi@0: 
aoqi@0:         //
aoqi@0:         // Since the tables are too ridiculous to use in code,
aoqi@0:         // we're using the footnotes here to drive this test.
aoqi@0:         //
aoqi@0:         switch (Character.getType(c)) {
aoqi@0:         // app. B footnote says these are 'name start'
aoqi@0:         // chars' ...
aoqi@0:         case Character.LOWERCASE_LETTER:        // Ll
aoqi@0:         case Character.UPPERCASE_LETTER:        // Lu
aoqi@0:         case Character.OTHER_LETTER:            // Lo
aoqi@0:         case Character.TITLECASE_LETTER:        // Lt
aoqi@0:         case Character.LETTER_NUMBER:            // Nl
aoqi@0:             // ... and these are name characters 'other
aoqi@0:             // than name start characters'
aoqi@0:         case Character.COMBINING_SPACING_MARK:    // Mc
aoqi@0:         case Character.ENCLOSING_MARK:        // Me
aoqi@0:         case Character.NON_SPACING_MARK:        // Mn
aoqi@0:         case Character.MODIFIER_LETTER:        // Lm
aoqi@0:         case Character.DECIMAL_DIGIT_NUMBER:        // Nd
aoqi@0: 
aoqi@0:             // OK, here we just have some exceptions to check...
aoqi@0:             return !isCompatibilityChar(c)
aoqi@0:                     // per "5.14 of Unicode", rule out some combiners
aoqi@0:                     && !(c >= 0x20dd && c <= 0x20e0);
aoqi@0: 
aoqi@0:         default:
aoqi@0:             // added a character ...
aoqi@0:             return c == 0x0387;
aoqi@0:         }
aoqi@0:     }
aoqi@0: 
aoqi@0:     private static boolean isDigit(char c) {
aoqi@0:         // [88] Digit ::= ...
aoqi@0: 
aoqi@0:         //
aoqi@0:         // java.lang.Character.isDigit is correct from the XML point
aoqi@0:         // of view except that it allows "fullwidth" digits.
aoqi@0:         //
aoqi@0:         return Character.isDigit(c)
aoqi@0:                 && !((c >= 0xff10) && (c <= 0xff19));
aoqi@0:     }
aoqi@0: 
aoqi@0:     private static boolean isExtender(char c) {
aoqi@0:         // [89] Extender ::= ...
aoqi@0:         return c == 0x00b7 || c == 0x02d0 || c == 0x02d1 || c == 0x0387
aoqi@0:                 || c == 0x0640 || c == 0x0e46 || c == 0x0ec6
aoqi@0:                 || c == 0x3005 || (c >= 0x3031 && c <= 0x3035)
aoqi@0:                 || (c >= 0x309d && c <= 0x309e)
aoqi@0:                 || (c >= 0x30fc && c <= 0x30fe)
aoqi@0:                 ;
aoqi@0:     }
aoqi@0: }