/*
* reserved comment block
* DO NOT REMOVE OR ALTER!
*/
/*
* Copyright 1999-2004 The Apache Software Foundation.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
/*
* $Id: CharInfo.java,v 1.2.4.1 2005/09/15 08:15:14 suresh_emailid Exp $
*/
/**
* This class provides services that tell if a character should have
* special treatement, such as entity reference substitution or normalization
* of a newline character. It also provides character to entity reference
* lookup.
*
* DEVELOPERS: See Known Issue in the constructor.
*
* @xsl.usage internal
*/
final class CharInfo
{
/** Given a character, lookup a String to output (e.g. a decorated entity reference). */
/**
* The name of the HTML entities file.
* If specified, the file will be resource loaded with the default class loader.
*/
/**
* The name of the XML entities file.
* If specified, the file will be resource loaded with the default class loader.
*/
/** The horizontal tab character, which the parser should always normalize. */
/** The linefeed character, which the parser should always normalize. */
/** The carriage return character, which the parser should always normalize. */
/** This flag is an optimization for HTML entities. It false if entities
* other than quot (34), amp (38), lt (60) and gt (62) are defined
* in the range 0 to 127.
* @xsl.usage internal
*/
final boolean onlyQuotAmpLtGt;
/** Copy the first 0,1 ... ASCII_MAX values into an array */
/** Array of values is faster access than a set of bits
* to quickly check ASCII characters in attribute values.
*/
/** Array of values is faster access than a set of bits
* to quickly check ASCII characters in text nodes.
*/
/** An array of bits to record if the character is in the set.
* Although information in this array is complete, the
* isSpecialAttrASCII array is used first because access to its values
* is common and faster.
*/
// 5 for 32 bit words, 6 for 64 bit words ...
/*
* This constant is used to shift an integer to quickly
* calculate which element its bit is stored in.
* 5 for 32 bit words (int) , 6 for 64 bit words (long)
*/
/*
* A mask to get the low order bits which are used to
* calculate the value of the bit within a given word,
* that will represent the presence of the integer in the
* set.
*
* 0x1F for 32 bit words (int),
* or 0x3F for 64 bit words (long)
*/
/*
* This is used for optimizing the lookup of bits representing
* the integers in the set. It is the index of the first element
* in the array array_of_bits[] that is not used.
*/
private int firstWordNotUsed;
/**
* Constructor that reads in a resource file that describes the mapping of
* characters to entity references.
* This constructor is private, just to force the use
* of the getCharInfo(entitiesResource) factory
*
* Resource files must be encoded in UTF-8 and can either be properties
* files with a .properties extension assumed. Alternatively, they can
* have the following form, with no particular extension assumed:
*
* <pre>
* # First char # is a comment
* Entity numericValue
* quot 34
* amp 38
* </pre>
*
* @param entitiesResource Name of properties or resource file that should
* be loaded, which describes that mapping of characters to entity
* references.
*/
{
this(entitiesResource, method, false);
}
{
boolean noExtraEntities = true;
// Make various attempts to interpret the parameter as a properties
// file or resource file, as follows:
//
// 1) attempt to load .properties file using ResourceBundle
// 2) try using the class loader to find the specified file a resource
// file
// 3) try treating the resource a URI
if (internal) {
try {
// Load entity property files by using PropertyResourceBundle,
// cause of security issure for applets
} catch (Exception e) {}
}
while (keys.hasMoreElements()){
if (extraEntity(code))
noExtraEntities = false;
}
} else {
// Load user specified resource file by using URL loading, it
// requires a valid URI as parameter
try {
if (internal) {
} else {
} else {
}
try {
} catch (Exception e) {}
}
}
throw new RuntimeException(
}
// Fix Bugzilla#4000: force reading in UTF-8
// This creates the de facto standard that Xalan's resource
// files must be encoded in UTF-8. This should work in all
// JVMs.
//
// %REVIEW% KNOWN ISSUE: IT FAILS IN MICROSOFT VJ++, which
// didn't implement the UTF-8 encoding. Theoretically, we should
// simply let it fail in that case, since the JVM is obviously
// broken if it doesn't support such a basic standard. But
// since there are still some users attempting to use VJ++ for
// development, we have dropped in a fallback which makes a
// second attempt using the platform's default encoding. In VJ++
// this is apparently ASCII, which is subset of UTF-8... and
// since the strings we'll be reading here are also primarily
// limited to the 7-bit ASCII range (at least, in English
// versions of Xalan), this should work well enough to keep us
// on the air until we're ready to officially decommit from
// VJ++.
try {
} catch (UnsupportedEncodingException e) {
}
continue;
}
if (index > 1) {
++index;
if (index > 0) {
}
if (extraEntity(code))
noExtraEntities = false;
}
}
}
} catch (Exception e) {
throw new RuntimeException(
new Object[] { entitiesResource,
e.toString(),
e.toString()}));
} finally {
try {
}
}
}
/* initialize the array isCleanTextASCII[] with a cache of values
* for use by ToStream.character(char[], int , int)
* and the array isSpecialTextASCII[] with the opposite values
* (all in the name of performance!)
*/
{
isCleanTextASCII[ch] = true;
isSpecialTextASCII[ch] = false;
}
else {
isCleanTextASCII[ch] = false;
isSpecialTextASCII[ch] = true;
}
// initialize the array with a cache of the BitSet values
for (int i=0; i<ASCII_MAX; i++)
isSpecialAttrASCII[i] = get(i);
/* Now that we've used get(ch) just above to initialize the
* two arrays we will change by adding a tab to the set of
* special chars for XML (but not HTML!).
* We do this because a tab is always a
* special character in an XML attribute,
* but only a special character in XML text
* if it has an entity defined for it.
* This is the reason for this delay.
*/
{
isSpecialAttrASCII[S_HORIZONAL_TAB] = true;
}
}
/**
* Defines a new character reference. The reference's name and value are
* supplied. Nothing happens if the character reference is already defined.
* <p>Unlike internal entities, character references are a string to single
* character mapping. They are used to map non-ASCII characters both on
* parsing and printing, primarily for HTML documents. '<amp;' is an
* example of a character reference.</p>
*
* @param name The entity's name
* @param value The entity's value
*/
{
}
/**
* Map a character to a String. For example given
* the character '>' this method would return the fully decorated
* entity name "<".
* Strings for entity references are loaded from a properties file,
* but additional mappings defined through calls to defineChar2String()
* are possible. Such entity reference mappings could be over-ridden.
*
* This is reusing a stored key object, in an effort to avoid
* heap activity. Unfortunately, that introduces a threading risk.
* Simplest fix for now is to make it a synchronized method, or to give
* up the reuse; I see very little performance difference between them.
* Long-term solution would be to replace the hashtable with a sparse array
* keyed directly from the character's integer value; see DTM's
* string pool for a related solution.
*
* @param value The character that should be resolved to
* a String, e.g. resolve '>' to "<".
*
* @return The String that the character is mapped to, or null if not found.
* @xsl.usage internal
*/
{
}
/**
* Tell if the character argument that is from
* an attribute value should have special treatment.
*
* @param value the value of a character that is in an attribute value
* @return true if the character should have any special treatment,
* such as when writing out attribute values,
* or entity references.
* @xsl.usage internal
*/
{
// for performance try the values in the boolean array first,
// this is faster access than the BitSet for common ASCII values
return isSpecialAttrASCII[value];
// rather than java.util.BitSet, our private
// implementation is faster (and less general).
}
/**
* Tell if the character argument that is from a
* text node should have special treatment.
*
* @param value the value of a character that is in a text node
* @return true if the character should have any special treatment,
* such as when writing out attribute values,
* or entity references.
* @xsl.usage internal
*/
{
// for performance try the values in the boolean array first,
// this is faster access than the BitSet for common ASCII values
return isSpecialTextASCII[value];
// rather than java.util.BitSet, our private
// implementation is faster (and less general).
}
/**
* This method is used to determine if an ASCII character in
* a text node (not an attribute value) is "clean".
* @param value the character to check (0 to 127).
* @return true if the character can go to the writer as-is
* @xsl.usage internal
*/
{
return isCleanTextASCII[value];
}
// In the future one might want to use the array directly and avoid
// the method call, but I think the JIT alreay inlines this well enough
// so don't do it (for now) - bjm
// public final boolean[] getASCIIClean()
// {
// return isCleanTextASCII;
// }
final boolean internal){
new PrivilegedAction() {
return new CharInfo(entitiesFileName,
});
}
/**
* Factory that reads in a resource file that describes the mapping of
* characters to entity references.
*
* Resource files must be encoded in UTF-8 and have a format like:
* <pre>
* # First char # is a comment
* Entity numericValue
* quot 34
* amp 38
* </pre>
* (Note: Why don't we just switch to .properties files? Oct-01 -sc)
*
* @param entitiesResource Name of entities resource file that should
* be loaded, which describes that mapping of characters to entity references.
* @param method the output method type, which should be one of "xml", "html", "text"...
*
* @xsl.usage internal
*/
{
return charInfo;
}
// try to load it internally - cache
try {
method, true);
return charInfo;
} catch (Exception e) {}
// try to load it externally - do not cache
try {
method, false);
} catch (Exception e) {}
} else {
try {
} catch (TransformerException te) {
throw new WrappedRuntimeException(te);
}
}
method, false);
}
/** Table of user-specified char infos. */
/**
* Returns the array element holding the bit value for the
* given integer
* @param i the integer that might be in the set of integers
*
*/
private static int arrayIndex(int i) {
return (i >> SHIFT_PER_WORD);
}
/**
* For a given integer in the set it returns the single bit
* value used within a given word that represents whether
* the integer is in the set or not.
*/
private static int bit(int i) {
return ret;
}
/**
* Creates a new empty set of integers (characters)
* @param max the maximum integer to be in the set.
*/
return arr;
}
/**
* Adds the integer (character) to the set of integers.
* @param i the integer to add to the set, valid values are
* 0, 1, 2 ... up to the maximum that was specified at
* the creation of the set.
*/
private final void set(int i) {
setASCIIdirty(i);
int j = (i >> SHIFT_PER_WORD); // this word is used
int k = j + 1;
if(firstWordNotUsed < k) // for optimization purposes.
firstWordNotUsed = k;
}
/**
* Return true if the integer (character)is in the set of integers.
*
* This implementation uses an array of integers with 32 bits per
* integer. If a bit is set to 1 the corresponding integer is
* in the set of integers.
*
* @param i an integer that is tested to see if it is the
* set of integers, or not.
*/
private final boolean get(int i) {
boolean in_the_set = false;
int j = (i >> SHIFT_PER_WORD); // wordIndex(i)
// an optimization here, ... a quick test to see
// if this integer is beyond any of the words in use
if(j < firstWordNotUsed)
in_the_set = (array_of_bits[j] &
(1 << (i & LOW_ORDER_BITMASK))
) != 0; // 0L for 64 bit words
return in_the_set;
}
// record if there are any entities other than
// quot, amp, lt, gt (probably user defined)
/**
* @return true if the entity
* @param code The value of the character that has an entity defined
* for it.
*/
{
boolean extra = false;
if (entityValue < 128)
{
switch (entityValue)
{
case 34 : // quot
case 38 : // amp
case 60 : // lt
case 62 : // gt
break;
default : // other entity in range 0 to 127
extra = true;
}
}
return extra;
}
/**
* If the character is a printable ASCII character then
* mark it as not clean and needing replacement with
* a String on output.
* @param ch
*/
private void setASCIIdirty(int j)
{
if (0 <= j && j < ASCII_MAX)
{
isCleanTextASCII[j] = false;
isSpecialTextASCII[j] = true;
}
}
/**
* If the character is a printable ASCII character then
* mark it as and not needing replacement with
* a String on output.
* @param ch
*/
private void setASCIIclean(int j)
{
if (0 <= j && j < ASCII_MAX)
{
isCleanTextASCII[j] = true;
isSpecialTextASCII[j] = false;
}
}
{
}
/**
* Simple class for fast lookup of char values, when used with
* hashtables. You can set the char, then use it as a key.
*
* This class is a copy of the one in com.sun.org.apache.xml.internal.utils.
* It exists to cut the serializers dependancy on that package.
*
* @xsl.usage internal
*/
{
/** String value */
private char m_char;
/**
* Constructor CharKey
*
* @param key char value of this object.
*/
{
}
/**
* Default constructor for a CharKey.
*
* @param key char value of this object.
*/
public CharKey()
{
}
/**
* Get the hash value of the character.
*
* @return hash value of the character.
*/
public final void setChar(char c)
{
m_char = c;
}
/**
* Get the hash value of the character.
*
* @return hash value of the character.
*/
public final int hashCode()
{
return (int)m_char;
}
/**
* Override of equals() for this object
*
* @param obj to compare to
*
* @return True if this object equals this string value
*/
{
}
}
}