aoqi@0: /*
aoqi@0:  * Copyright (c) 1997, 2011, 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;
aoqi@0: 
aoqi@0: /**
aoqi@0:  * Processes white space normalization.
aoqi@0:  *
aoqi@0:  * @since 1.0
aoqi@0:  */
aoqi@0: public abstract class WhiteSpaceProcessor {
aoqi@0: 
aoqi@0: // benchmarking (see test/src/ReplaceTest.java in the CVS Attic)
aoqi@0: // showed that this code is slower than the current code.
aoqi@0: //
aoqi@0: //    public static String replace(String text) {
aoqi@0: //        final int len = text.length();
aoqi@0: //        StringBuffer result = new StringBuffer(len);
aoqi@0: //
aoqi@0: //        for (int i = 0; i < len; i++) {
aoqi@0: //            char ch = text.charAt(i);
aoqi@0: //            if (isWhiteSpace(ch))
aoqi@0: //                result.append(' ');
aoqi@0: //            else
aoqi@0: //                result.append(ch);
aoqi@0: //        }
aoqi@0: //
aoqi@0: //        return result.toString();
aoqi@0: //    }
aoqi@0: 
aoqi@0:     public static String replace(String text) {
aoqi@0:         return replace( (CharSequence)text ).toString();
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * @since 2.0
aoqi@0:      */
aoqi@0:     public static CharSequence replace(CharSequence text) {
aoqi@0:         int i=text.length()-1;
aoqi@0: 
aoqi@0:         // look for the first whitespace char.
aoqi@0:         while( i>=0 && !isWhiteSpaceExceptSpace(text.charAt(i)) )
aoqi@0:             i--;
aoqi@0: 
aoqi@0:         if( i<0 )
aoqi@0:             // no such whitespace. replace(text)==text.
aoqi@0:             return text;
aoqi@0: 
aoqi@0:         // we now know that we need to modify the text.
aoqi@0:         // allocate a char array to do it.
aoqi@0:         StringBuilder buf = new StringBuilder(text);
aoqi@0: 
aoqi@0:         buf.setCharAt(i--,' ');
aoqi@0:         for( ; i>=0; i-- )
aoqi@0:             if( isWhiteSpaceExceptSpace(buf.charAt(i)))
aoqi@0:                 buf.setCharAt(i,' ');
aoqi@0: 
aoqi@0:         return new String(buf);
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Equivalent of {@link String#trim()}.
aoqi@0:      * @since 2.0
aoqi@0:      */
aoqi@0:     public static CharSequence trim(CharSequence text) {
aoqi@0:         int len = text.length();
aoqi@0:         int start = 0;
aoqi@0: 
aoqi@0:         while( start<len && isWhiteSpace(text.charAt(start)) )
aoqi@0:             start++;
aoqi@0: 
aoqi@0:         int end = len-1;
aoqi@0: 
aoqi@0:         while( end>start && isWhiteSpace(text.charAt(end)) )
aoqi@0:             end--;
aoqi@0: 
aoqi@0:         if(start==0 && end==len-1)
aoqi@0:             return text;    // no change
aoqi@0:         else
aoqi@0:             return text.subSequence(start,end+1);
aoqi@0:     }
aoqi@0: 
aoqi@0:     public static String collapse(String text) {
aoqi@0:         return collapse( (CharSequence)text ).toString();
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * This is usually the biggest processing bottleneck.
aoqi@0:      *
aoqi@0:      * @since 2.0
aoqi@0:      */
aoqi@0:     public static CharSequence collapse(CharSequence text) {
aoqi@0:         int len = text.length();
aoqi@0: 
aoqi@0:         // most of the texts are already in the collapsed form.
aoqi@0:         // so look for the first whitespace in the hope that we will
aoqi@0:         // never see it.
aoqi@0:         int s=0;
aoqi@0:         while(s<len) {
aoqi@0:             if(isWhiteSpace(text.charAt(s)))
aoqi@0:                 break;
aoqi@0:             s++;
aoqi@0:         }
aoqi@0:         if(s==len)
aoqi@0:             // the input happens to be already collapsed.
aoqi@0:             return text;
aoqi@0: 
aoqi@0:         // we now know that the input contains spaces.
aoqi@0:         // let's sit down and do the collapsing normally.
aoqi@0: 
aoqi@0:         StringBuilder result = new StringBuilder(len /*allocate enough size to avoid re-allocation*/ );
aoqi@0: 
aoqi@0:         if(s!=0) {
aoqi@0:             for( int i=0; i<s; i++ )
aoqi@0:                 result.append(text.charAt(i));
aoqi@0:             result.append(' ');
aoqi@0:         }
aoqi@0: 
aoqi@0:         boolean inStripMode = true;
aoqi@0:         for (int i = s+1; i < len; i++) {
aoqi@0:             char ch = text.charAt(i);
aoqi@0:             boolean b = isWhiteSpace(ch);
aoqi@0:             if (inStripMode && b)
aoqi@0:                 continue; // skip this character
aoqi@0: 
aoqi@0:             inStripMode = b;
aoqi@0:             if (inStripMode)
aoqi@0:                 result.append(' ');
aoqi@0:             else
aoqi@0:                 result.append(ch);
aoqi@0:         }
aoqi@0: 
aoqi@0:         // remove trailing whitespaces
aoqi@0:         len = result.length();
aoqi@0:         if (len > 0 && result.charAt(len - 1) == ' ')
aoqi@0:             result.setLength(len - 1);
aoqi@0:         // whitespaces are already collapsed,
aoqi@0:         // so all we have to do is to remove the last one character
aoqi@0:         // if it's a whitespace.
aoqi@0: 
aoqi@0:         return result;
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Returns true if the specified string is all whitespace.
aoqi@0:      */
aoqi@0:     public static boolean isWhiteSpace(CharSequence s) {
aoqi@0:         for( int i=s.length()-1; i>=0; i-- )
aoqi@0:             if(!isWhiteSpace(s.charAt(i)))
aoqi@0:                 return false;
aoqi@0:         return true;
aoqi@0:     }
aoqi@0: 
aoqi@0:     /** returns true if the specified char is a white space character. */
aoqi@0:     public static boolean isWhiteSpace(char ch) {
aoqi@0:         // most of the characters are non-control characters.
aoqi@0:         // so check that first to quickly return false for most of the cases.
aoqi@0:         if( ch>0x20 )   return false;
aoqi@0: 
aoqi@0:         // other than we have to do four comparisons.
aoqi@0:         return ch == 0x9 || ch == 0xA || ch == 0xD || ch == 0x20;
aoqi@0:     }
aoqi@0: 
aoqi@0:     /**
aoqi@0:      * Returns true if the specified char is a white space character
aoqi@0:      * but not 0x20.
aoqi@0:      */
aoqi@0:     protected static boolean isWhiteSpaceExceptSpace(char ch) {
aoqi@0:         // most of the characters are non-control characters.
aoqi@0:         // so check that first to quickly return false for most of the cases.
aoqi@0:         if( ch>=0x20 )   return false;
aoqi@0: 
aoqi@0:         // other than we have to do four comparisons.
aoqi@0:         return ch == 0x9 || ch == 0xA || ch == 0xD;
aoqi@0:     }
aoqi@0: }