aoqi@0: /*
aoqi@0:  * Copyright (c) 1997, 2010, 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.bind.api.impl;
aoqi@0: 
aoqi@0: import javax.lang.model.SourceVersion;
aoqi@0: import java.util.ArrayList;
aoqi@0: import java.util.List;
aoqi@0: import java.util.StringTokenizer;
aoqi@0: 
aoqi@0: /**
aoqi@0:  * Converts aribitrary strings into Java identifiers.
aoqi@0:  *
aoqi@0:  * @author
aoqi@0:  *    <a href="mailto:kohsuke.kawaguchi@sun.com">Kohsuke KAWAGUCHI</a>
aoqi@0:  */
aoqi@0: public interface NameConverter
aoqi@0: {
aoqi@0:     /**
aoqi@0:      * converts a string into an identifier suitable for classes.
aoqi@0:      *
aoqi@0:      * In general, this operation should generate "NamesLikeThis".
aoqi@0:      */
aoqi@0:     String toClassName( String token );
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * converts a string into an identifier suitable for interfaces.
aoqi@0:      *
aoqi@0:      * In general, this operation should generate "NamesLikeThis".
aoqi@0:      * But for example, it can prepend every interface with 'I'.
aoqi@0:      */
aoqi@0:     String toInterfaceName( String token );
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * converts a string into an identifier suitable for properties.
aoqi@0:      *
aoqi@0:      * In general, this operation should generate "NamesLikeThis",
aoqi@0:      * which will be used with known prefixes like "get" or "set".
aoqi@0:      */
aoqi@0:     String toPropertyName( String token );
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * converts a string into an identifier suitable for constants.
aoqi@0:      *
aoqi@0:      * In the standard Java naming convention, this operation should
aoqi@0:      * generate "NAMES_LIKE_THIS".
aoqi@0:      */
aoqi@0:     String toConstantName( String token );
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Converts a string into an identifier suitable for variables.
aoqi@0:      *
aoqi@0:      * In general it should generate "namesLikeThis".
aoqi@0:      */
aoqi@0:     String toVariableName( String token );
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Converts a namespace URI into a package name.
aoqi@0:      * This method should expect strings like
aoqi@0:      * "http://foo.bar.zot/org", "urn:abc:def:ghi" "", or even "###"
aoqi@0:      * (basically anything) and expected to return a package name,
aoqi@0:      * liks "org.acme.foo".
aoqi@0:      *
aoqi@0:      */
aoqi@0:     String toPackageName( String namespaceUri );
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * The name converter implemented by Code Model.
aoqi@0:      *
aoqi@0:      * This is the standard name conversion for JAXB.
aoqi@0:      */
aoqi@0:     public static final NameConverter standard = new Standard();
aoqi@0: 
aoqi@0:     static class Standard extends NameUtil implements NameConverter {
aoqi@0:         public String toClassName(String s) {
aoqi@0:             return toMixedCaseName(toWordList(s), true);
aoqi@0:         }
aoqi@0:         public String toVariableName(String s) {
aoqi@0:             return toMixedCaseName(toWordList(s), false);
aoqi@0:         }
aoqi@0:         public String toInterfaceName( String token ) {
aoqi@0:             return toClassName(token);
aoqi@0:         }
aoqi@0:         public String toPropertyName(String s) {
aoqi@0:             String prop = toClassName(s);
aoqi@0:             // property name "Class" with collide with Object.getClass,
aoqi@0:             // so escape this.
aoqi@0:             if(prop.equals("Class"))
aoqi@0:                 prop = "Clazz";
aoqi@0:             return prop;
aoqi@0:         }
aoqi@0:         public String toConstantName( String token ) {
aoqi@0:             return super.toConstantName(token);
aoqi@0:         }
aoqi@0:         /**
aoqi@0:          * Computes a Java package name from a namespace URI,
aoqi@0:          * as specified in the spec.
aoqi@0:          *
aoqi@0:          * @return
aoqi@0:          *      null if it fails to derive a package name.
aoqi@0:          */
aoqi@0:         public String toPackageName( String nsUri ) {
aoqi@0:             // remove scheme and :, if present
aoqi@0:             // spec only requires us to remove 'http' and 'urn'...
aoqi@0:             int idx = nsUri.indexOf(':');
aoqi@0:             String scheme = "";
aoqi@0:             if(idx>=0) {
aoqi@0:                 scheme = nsUri.substring(0,idx);
aoqi@0:                 if( scheme.equalsIgnoreCase("http") || scheme.equalsIgnoreCase("urn") )
aoqi@0:                     nsUri = nsUri.substring(idx+1);
aoqi@0:             }
aoqi@0: 
aoqi@0:             // tokenize string
aoqi@0:             ArrayList<String> tokens = tokenize( nsUri, "/: " );
aoqi@0:             if( tokens.size() == 0 ) {
aoqi@0:                 return null;
aoqi@0:             }
aoqi@0: 
aoqi@0:             // remove trailing file type, if necessary
aoqi@0:             if( tokens.size() > 1 ) {
aoqi@0:                 // for uri's like "www.foo.com" and "foo.com", there is no trailing
aoqi@0:                 // file, so there's no need to look at the last '.' and substring
aoqi@0:                 // otherwise, we loose the "com" (which would be wrong)
aoqi@0:                 String lastToken = tokens.get( tokens.size()-1 );
aoqi@0:                 idx = lastToken.lastIndexOf( '.' );
aoqi@0:                 if( idx > 0 ) {
aoqi@0:                     lastToken = lastToken.substring( 0, idx );
aoqi@0:                     tokens.set( tokens.size()-1, lastToken );
aoqi@0:                 }
aoqi@0:             }
aoqi@0: 
aoqi@0:             // tokenize domain name and reverse.  Also remove :port if it exists
aoqi@0:             String domain = tokens.get( 0 );
aoqi@0:             idx = domain.indexOf(':');
aoqi@0:             if( idx >= 0) domain = domain.substring(0, idx);
aoqi@0:             ArrayList<String> r = reverse( tokenize( domain, scheme.equals("urn")?".-":"." ) );
aoqi@0:             if( r.get( r.size()-1 ).equalsIgnoreCase( "www" ) ) {
aoqi@0:                 // remove leading www
aoqi@0:                 r.remove( r.size()-1 );
aoqi@0:             }
aoqi@0: 
aoqi@0:             // replace the domain name with tokenized items
aoqi@0:             tokens.addAll( 1, r );
aoqi@0:             tokens.remove( 0 );
aoqi@0: 
aoqi@0:             // iterate through the tokens and apply xml->java name algorithm
aoqi@0:             for( int i = 0; i < tokens.size(); i++ ) {
aoqi@0: 
aoqi@0:                 // get the token and remove illegal chars
aoqi@0:                 String token = tokens.get( i );
aoqi@0:                 token = removeIllegalIdentifierChars( token );
aoqi@0: 
aoqi@0:                 // this will check for reserved keywords
aoqi@0:                 if (SourceVersion.isKeyword(token.toLowerCase())) {
aoqi@0:                     token = '_' + token;
aoqi@0:                 }
aoqi@0: 
aoqi@0:                 tokens.set( i, token.toLowerCase() );
aoqi@0:             }
aoqi@0: 
aoqi@0:             // concat all the pieces and return it
aoqi@0:             return combine( tokens, '.' );
aoqi@0:         }
aoqi@0: 
aoqi@0: 
aoqi@0:         private static String removeIllegalIdentifierChars(String token) {
aoqi@0:             StringBuilder newToken = new StringBuilder(token.length() + 1); // max expected length
aoqi@0:             for( int i = 0; i < token.length(); i++ ) {
aoqi@0:                 char c = token.charAt( i );
aoqi@0:                 if (i == 0 && !Character.isJavaIdentifierStart(c)) { // c can't be used as FIRST char
aoqi@0:                     newToken.append('_');
aoqi@0:                 }
aoqi@0:                 if (!Character.isJavaIdentifierPart(c)) { // c can't be used
aoqi@0:                     newToken.append('_');
aoqi@0:                 } else {
aoqi@0:                     newToken.append(c); // c is valid
aoqi@0:                 }
aoqi@0:             }
aoqi@0:             return newToken.toString();
aoqi@0:         }
aoqi@0: 
aoqi@0: 
aoqi@0:         private static ArrayList<String> tokenize( String str, String sep ) {
aoqi@0:             StringTokenizer tokens = new StringTokenizer(str,sep);
aoqi@0:             ArrayList<String> r = new ArrayList<String>();
aoqi@0: 
aoqi@0:             while(tokens.hasMoreTokens())
aoqi@0:                 r.add( tokens.nextToken() );
aoqi@0: 
aoqi@0:             return r;
aoqi@0:         }
aoqi@0: 
aoqi@0:         private static <T> ArrayList<T> reverse( List<T> a ) {
aoqi@0:             ArrayList<T> r = new ArrayList<T>();
aoqi@0: 
aoqi@0:             for( int i=a.size()-1; i>=0; i-- )
aoqi@0:                 r.add( a.get(i) );
aoqi@0: 
aoqi@0:             return r;
aoqi@0:         }
aoqi@0: 
aoqi@0:         private static String combine( List r, char sep ) {
aoqi@0:             StringBuilder buf = new StringBuilder(r.get(0).toString());
aoqi@0: 
aoqi@0:             for( int i=1; i<r.size(); i++ ) {
aoqi@0:                 buf.append(sep);
aoqi@0:                 buf.append(r.get(i));
aoqi@0:             }
aoqi@0: 
aoqi@0:             return buf.toString();
aoqi@0:         }
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * JAX-PRC compatible name converter implementation.
aoqi@0:      *
aoqi@0:      * The only difference is that we treat '_' as a valid character
aoqi@0:      * and not as a word separator.
aoqi@0:      */
aoqi@0:     public static final NameConverter jaxrpcCompatible = new Standard() {
aoqi@0:         protected boolean isPunct(char c) {
aoqi@0:             return (c == '.' || c == '-' || c == ';' /*|| c == '_'*/ || c == '\u00b7'
aoqi@0:                     || c == '\u0387' || c == '\u06dd' || c == '\u06de');
aoqi@0:         }
aoqi@0:         protected boolean isLetter(char c) {
aoqi@0:             return super.isLetter(c) || c=='_';
aoqi@0:         }
aoqi@0: 
aoqi@0:         protected int classify(char c0) {
aoqi@0:             if(c0=='_') return NameUtil.OTHER_LETTER;
aoqi@0:             return super.classify(c0);
aoqi@0:         }
aoqi@0:     };
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Smarter converter used for RELAX NG support.
aoqi@0:      */
aoqi@0:     public static final NameConverter smart = new Standard() {
aoqi@0:         public String toConstantName( String token ) {
aoqi@0:             String name = super.toConstantName(token);
aoqi@0:             if(!SourceVersion.isKeyword(name))
aoqi@0:                 return name;
aoqi@0:             else
aoqi@0:                 return '_'+name;
aoqi@0:         }
aoqi@0:     };
aoqi@0: }