icu4jsrc

[Dictionary.git] / jars / icu4j-4_2_1-src / src / com / ibm / icu / impl / ICUResourceBundleReader.java

/*\r
 *******************************************************************************\r
 * Copyright (C) 2004-2008, International Business Machines Corporation and         *\r
 * others. All Rights Reserved.                                                *\r
 *******************************************************************************\r
 */\r
package com.ibm.icu.impl;\r
\r
import java.io.BufferedInputStream;\r
import java.io.DataInputStream;\r
import java.io.IOException;\r
import java.io.InputStream;\r
\r
import com.ibm.icu.util.ULocale;\r
import com.ibm.icu.util.VersionInfo;\r
\r
\r
\r
/**\r
 * This class reads the *.res resource bundle format\r
 *\r
 * (For the latest version of the file format documentation see\r
 * ICU4C's source/common/uresdata.h file.)\r
 *\r
 * File format for .res resource bundle files (formatVersion=1.2)\r
 *\r
 * An ICU4C resource bundle file (.res) is a binary, memory-mappable file\r
 * with nested, hierarchical data structures.\r
 * It physically contains the following:\r
 *\r
 *   Resource root; -- 32-bit Resource item, root item for this bundle's tree;\r
 *                     currently, the root item must be a table or table32 resource item\r
 *   int32_t indexes[indexes[0]]; -- array of indexes for friendly\r
 *                                   reading and swapping; see URES_INDEX_* above\r
 *                                   new in formatVersion 1.1 (ICU 2.8)\r
 *   char keys[]; -- characters for key strings\r
 *                   (formatVersion 1.0: up to 65k of characters; 1.1: <2G)\r
 *                   (minus the space for root and indexes[]),\r
 *                   which consist of invariant characters (ASCII/EBCDIC) and are NUL-terminated;\r
 *                   padded to multiple of 4 bytes for 4-alignment of the following data\r
 *   data; -- data directly and indirectly indexed by the root item;\r
 *            the structure is determined by walking the tree\r
 *\r
 * Each resource bundle item has a 32-bit Resource handle (see typedef above)\r
 * which contains the item type number in its upper 4 bits (31..28) and either\r
 * an offset or a direct value in its lower 28 bits (27..0).\r
 * The order of items is undefined and only determined by walking the tree.\r
 * Leaves of the tree may be stored first or last or anywhere in between,\r
 * and it is in theory possible to have unreferenced holes in the file.\r
 *\r
 * Direct values:\r
 * - Empty Unicode strings have an offset value of 0 in the Resource handle itself.\r
 * - Integer values are 28-bit values stored in the Resource handle itself;\r
 *   the interpretation of unsigned vs. signed integers is up to the application.\r
 *\r
 * All other types and values use 28-bit offsets to point to the item's data.\r
 * The offset is an index to the first 32-bit word of the value, relative to the\r
 * start of the resource data (i.e., the root item handle is at offset 0).\r
 * To get byte offsets, the offset is multiplied by 4 (or shifted left by 2 bits).\r
 * All resource item values are 4-aligned.\r
 *\r
 * The structures (memory layouts) for the values for each item type are listed\r
 * in the table above.\r
 *\r
 * Nested, hierarchical structures: -------------\r
 *\r
 * Table items contain key-value pairs where the keys are 16-bit offsets to char * key strings.\r
 * Key string offsets are also relative to the start of the resource data (of the root handle),\r
 * i.e., the first string has an offset of 4 (after the 4-byte root handle).\r
 *\r
 * The values of these pairs are Resource handles.\r
 *\r
 * Array items are simple vectors of Resource handles.\r
 *\r
 * An alias item is special (and new in ICU 2.4): --------------\r
 *\r
 * Its memory layout is just like for a UnicodeString, but at runtime it resolves to\r
 * another resource bundle's item according to the path in the string.\r
 * This is used to share items across bundles that are in different lookup/fallback\r
 * chains (e.g., large collation data among zh_TW and zh_HK).\r
 * This saves space (for large items) and maintenance effort (less duplication of data).\r
 *\r
 * --------------------------------------------------------------------------\r
 *\r
 * Resource types:\r
 *\r
 * Most resources have their values stored at four-byte offsets from the start\r
 * of the resource data. These values are at least 4-aligned.\r
 * Some resource values are stored directly in the offset field of the Resource itself.\r
 * See UResType in unicode/ures.h for enumeration constants for Resource types.\r
 *\r
 * Type Name            Memory layout of values\r
 *                      (in parentheses: scalar, non-offset values)\r
 *\r
 * 0  Unicode String:   int32_t length, UChar[length], (UChar)0, (padding)\r
 *                  or  (empty string ("") if offset==0)\r
 * 1  Binary:           int32_t length, uint8_t[length], (padding)\r
 *                      - this value should be 32-aligned -\r
 * 2  Table:            uint16_t count, uint16_t keyStringOffsets[count], (uint16_t padding), Resource[count]\r
 * 3  Alias:            (physically same value layout as string, new in ICU 2.4)\r
 * 4  Table32:          int32_t count, int32_t keyStringOffsets[count], Resource[count]\r
 *                      (new in formatVersion 1.1/ICU 2.8)\r
 *\r
 * 7  Integer:          (28-bit offset is integer value)\r
 * 8  Array:            int32_t count, Resource[count]\r
 *\r
 * 14 Integer Vector:   int32_t length, int32_t[length]\r
 * 15 Reserved:         This value denotes special purpose resources and is for internal use.\r
 *\r
 * Note that there are 3 types with data vector values:\r
 * - Vectors of 8-bit bytes stored as type Binary.\r
 * - Vectors of 16-bit words stored as type Unicode String\r
 *                     (no value restrictions, all values 0..ffff allowed!).\r
 * - Vectors of 32-bit words stored as type Integer Vector.\r
 *\r
 *\r
 */\r
public final class ICUResourceBundleReader implements ICUBinary.Authenticate{\r
\r
    /**\r
     * File format version that this class understands.\r
     * "ResB"\r
     */\r
    private static final byte DATA_FORMAT_ID[] = {(byte)0x52, (byte)0x65, \r
                                                     (byte)0x73, (byte)0x42};\r
\r
    private static final String ICU_RESOURCE_SUFFIX = ".res";\r
    \r
    /* indexes[] value names; indexes are generally 32-bit (Resource) indexes */\r
    private static final int    URES_INDEX_LENGTH           = 0;        /* [0] contains URES_INDEX_TOP==the length of indexes[] */\r
    //private static final int    URES_INDEX_STRINGS_TOP      = 1;        /* [1] contains the top of the strings, */\r
                                                                        /*     same as the bottom of resources, rounded up */\r
    //private static final int    URES_INDEX_RESOURCES_TOP    = 2;        /* [2] contains the top of all resources */\r
    private static final int    URES_INDEX_BUNDLE_TOP       = 3;        /* [3] contains the top of the bundle, */\r
                                                                        /*     in case it were ever different from [2] */\r
    //private static final int    URES_INDEX_MAX_TABLE_LENGTH = 4;        /* [4] max. length of any table */\r
    private static final int    URES_INDEX_ATTRIBUTES       = 5;        /* [5] attributes bit set, see URES_ATT_* (new in formatVersion 1.2) */\r
    //private static final int    URES_INDEX_TOP              = 6;\r
\r
    //private static final int    URES_STRINGS_BOTTOM=(1+URES_INDEX_TOP)*4;\r
\r
    /*\r
     * Nofallback attribute, attribute bit 0 in indexes[URES_INDEX_ATTRIBUTES].\r
     * New in formatVersion 1.2 (ICU 3.6).\r
     *\r
     * If set, then this resource bundle is a standalone bundle.\r
     * If not set, then the bundle participates in locale fallback, eventually\r
     * all the way to the root bundle.\r
     * If indexes[] is missing or too short, then the attribute cannot be determined\r
     * reliably. Dependency checking should ignore such bundles, and loading should\r
     * use fallbacks.\r
     */\r
    private static final int URES_ATT_NO_FALLBACK = 1;\r
\r
    private static final boolean DEBUG = false;\r
    \r
    private byte[] /* formatVersion, */ dataVersion;\r
\r
    private int rootRes;\r
    private int[] indexes;\r
    private boolean noFallback; /* see URES_ATT_NO_FALLBACK */\r
\r
    private byte[] data;\r
\r
    private ICUResourceBundleReader(InputStream stream, String resolvedName){\r
\r
        BufferedInputStream bs = new BufferedInputStream(stream);\r
        try{\r
            if(DEBUG) System.out.println("The InputStream class is: " + stream.getClass().getName());\r
            if(DEBUG) System.out.println("The BufferedInputStream class is: " + bs.getClass().getName());\r
            if(DEBUG) System.out.println("The bytes avialable in stream before reading the header: " + bs.available());\r
            \r
            dataVersion = ICUBinary.readHeader(bs,DATA_FORMAT_ID,this);\r
\r
            if(DEBUG) System.out.println("The bytes available in stream after reading the header: " + bs.available());\r
                 \r
            readData(bs);\r
            stream.close();\r
        }catch(IOException ex){\r
            throw new RuntimeException("Data file "+ resolvedName+ " is corrupt - " + ex.getMessage());   \r
        }\r
    }\r
    public static ICUResourceBundleReader getReader(String baseName, String localeName, ClassLoader root){\r
        String resolvedName = getFullName(baseName, localeName);\r
        InputStream stream = ICUData.getStream(root,resolvedName);\r
        \r
        if(stream==null){\r
            return null;\r
        }\r
        ICUResourceBundleReader reader = new ICUResourceBundleReader(stream, resolvedName);\r
        return reader;\r
    }\r
\r
    private static void writeInt(int i, byte[] bytes, int offset) {\r
        bytes[offset++]=(byte)(i>>24);\r
        bytes[offset++]=(byte)(i>>16);\r
        bytes[offset++]=(byte)(i>>8);\r
        bytes[offset]=(byte)i;\r
    }\r
\r
    private void readData(InputStream stream)\r
            throws IOException{\r
        \r
        DataInputStream ds = new DataInputStream(stream);\r
\r
        if(DEBUG) System.out.println("The DataInputStream class is: " + ds.getClass().getName());\r
        if(DEBUG) System.out.println("The available bytes in the stream before reading the data: "+ds.available());\r
\r
        /*\r
         * The following will read two integers before ds.mark().\r
         * Later, the two integers need to be placed into data[],\r
         * then ds.reset(), then ds.readFully(into rest of data[]).\r
         *\r
         * This is necessary because we don't know the readLimit for ds.mark()\r
         * until we have read the second integer (indexLength).\r
         */\r
        rootRes = ds.readInt();\r
\r
        // read the variable-length indexes[] array\r
        int indexLength = ds.readInt();\r
        ds.mark((indexLength-1)*4);\r
\r
        indexes = new int[indexLength];\r
        indexes[URES_INDEX_LENGTH] = indexLength;\r
\r
        for(int i=1; i<indexLength; i++){\r
            indexes[i] = ds.readInt();   \r
        }\r
\r
        // determine if this resource bundle falls back to a parent bundle\r
        // along normal locale ID fallback\r
        noFallback =\r
            indexLength > URES_INDEX_ATTRIBUTES &&\r
            (indexes[URES_INDEX_ATTRIBUTES]&URES_ATT_NO_FALLBACK)!=0;\r
\r
        // read the entire bundle (after the header) into data[]\r
        // put rootRes and indexLength into data[0..7]\r
        // and the rest of the data into data[8..length-1]\r
        int length = indexes[URES_INDEX_BUNDLE_TOP]*4;\r
        if(DEBUG) System.out.println("The number of bytes in the bundle: "+length);\r
    \r
        data = new byte[length];\r
        writeInt(rootRes, data, 0);\r
        writeInt(indexLength, data, 4);\r
\r
        // now reset to the mark, which was set after reading rootRes and indexLength\r
        ds.reset();\r
        ds.readFully(data, 8, length-8);\r
    }\r
\r
    /**\r
     * Gets the full name of the resource with suffix.\r
     */\r
    public static String getFullName(String baseName, String localeName){\r
        if(baseName==null || baseName.length()==0){\r
            if(localeName.length()==0){\r
                return ULocale.getDefault().toString()+ICU_RESOURCE_SUFFIX;   \r
            }else{\r
                return localeName+ICU_RESOURCE_SUFFIX;\r
            }\r
        }else{\r
            if(baseName.indexOf('.')==-1){\r
                if(baseName.charAt(baseName.length()-1)!= '/'){\r
                    return baseName+"/"+localeName+ICU_RESOURCE_SUFFIX;\r
                }else{\r
                    return baseName+localeName+ICU_RESOURCE_SUFFIX;   \r
                }\r
            }else{\r
                baseName = baseName.replace('.','/');\r
                if(localeName.length()==0){\r
                    return baseName+ICU_RESOURCE_SUFFIX;   \r
                }else{\r
                    return baseName+"_"+localeName+ICU_RESOURCE_SUFFIX;\r
                }\r
            }\r
        }\r
    }\r
    \r
    public VersionInfo getVersion(){\r
        return VersionInfo.getInstance(dataVersion[0],dataVersion[1],dataVersion[2],dataVersion[3]);   \r
    }\r
    public boolean isDataVersionAcceptable(byte version[]){\r
        // while ICU4C can read formatVersion 1.0 and up,\r
        // ICU4J requires 1.1 as a minimum\r
        // formatVersion = version;\r
        return version[0] == 1 && version[1] >= 1;\r
    }\r
    \r
    public byte[] getData(){\r
        return data;   \r
    }\r
    public int getRootResource() {\r
        return rootRes;\r
    }\r
    public boolean getNoFallback() {\r
        return noFallback;\r
    }\r
}\r