sal/textenc/tenchelp.hxx

   1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
   2 /*************************************************************************
   3  *
   4  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   5  *
   6  * Copyright 2000, 2010 Oracle and/or its affiliates.
   7  *
   8  * OpenOffice.org - a multi-platform office productivity suite
   9  *
  10  * This file is part of OpenOffice.org.
  11  *
  12  * OpenOffice.org is free software: you can redistribute it and/or modify
  13  * it under the terms of the GNU Lesser General Public License version 3
  14  * only, as published by the Free Software Foundation.
  15  *
  16  * OpenOffice.org is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19  * GNU Lesser General Public License version 3 for more details
  20  * (a copy is included in the LICENSE file that accompanied this code).
  21  *
  22  * You should have received a copy of the GNU Lesser General Public License
  23  * version 3 along with OpenOffice.org.  If not, see
  24  * <http://www.openoffice.org/license.html>
  25  * for a copy of the LGPLv3 License.
  26  *
  27  ************************************************************************/
  28
  29 #ifndef INCLUDED_SAL_TEXTENC_TENCHELP_HXX
  30 #define INCLUDED_SAL_TEXTENC_TENCHELP_HXX
  31
  32 #include "sal/config.h"
  33
  34 #include "rtl/tencinfo.h"
  35 #include "rtl/textcvt.h"
  36 #include "rtl/textenc.h"
  37 #include "sal/types.h"
  38
  39 #include "unichars.hxx"
  40
  41 #define RTL_TEXTCVT_BYTE_PRIVATE_START 0xF100
  42 #define RTL_TEXTCVT_BYTE_PRIVATE_END 0xF1FF
  43
  44 /* ----------------- */
  45 /* - TextConverter - */
  46 /* ----------------- */
  47
  48 typedef
  49 sal_Size (* ImplConvertToUnicodeProc)(void const * pData,
  50                                       void * pContext,
  51                                       char const * pSrcBuf,
  52                                       sal_Size nSrcBytes,
  53                                       sal_Unicode * pDestBuf,
  54                                       sal_Size nDestChars,
  55                                       sal_uInt32 nFlags,
  56                                       sal_uInt32 * pInfo,
  57                                       sal_Size * pSrcCvtBytes);
  58
  59 typedef
  60 sal_Size (* ImplConvertToTextProc)(void const * pData,
  61                                    void * pContext,
  62                                    sal_Unicode const * pSrcBuf,
  63                                    sal_Size nSrcChars,
  64                                    char * pDestBuf,
  65                                    sal_Size nDestBytes,
  66                                    sal_uInt32 nFlags,
  67                                    sal_uInt32 * pInfo,
  68                                    sal_Size * pSrcCvtChars);
  69
  70 typedef void * (* ImplCreateTextContextProc)();
  71
  72 typedef void (* ImplDestroyTextContextProc)(void * pContext);
  73
  74 typedef void (* ImplResetTextContextProc)(void * pContext);
  75
  76 typedef void * (* ImplCreateUnicodeContextProc)();
  77
  78 typedef void (* ImplDestroyUnicodeContextProc)(void * pContext);
  79
  80 typedef void (* ImplResetUnicodeContextProc)(void * pContext);
  81
  82 struct ImplTextConverter
  83 {
  84     void const * mpConvertData;
  85     ImplConvertToUnicodeProc mpConvertTextToUnicodeProc;
  86     ImplConvertToTextProc mpConvertUnicodeToTextProc;
  87     ImplCreateTextContextProc mpCreateTextToUnicodeContext;
  88     ImplDestroyTextContextProc mpDestroyTextToUnicodeContext;
  89     ImplResetTextContextProc mpResetTextToUnicodeContext;
  90     ImplCreateUnicodeContextProc mpCreateUnicodeToTextContext;
  91     ImplDestroyUnicodeContextProc mpDestroyUnicodeToTextContext;
  92     ImplResetUnicodeContextProc mpResetUnicodeToTextContext;
  93 };
  94
  95 /* ----------------------------- */
  96 /* - TextEncoding - Structures - */
  97 /* ----------------------------- */
  98
  99 struct ImplTextEncodingData
 100 {
 101     ImplTextConverter maConverter;
 102     sal_uInt8 mnMinCharSize;
 103     sal_uInt8 mnMaxCharSize;
 104     sal_uInt8 mnAveCharSize;
 105     sal_uInt8 mnBestWindowsCharset;
 106     char const * mpBestUnixCharset;
 107     char const * mpBestMimeCharset;
 108     sal_uInt32 mnInfoFlags;
 109 };
 110
 111 /* ----------------------------------- */
 112 /* - TextConverter - Byte-Structures - */
 113 /* ----------------------------------- */
 114
 115 struct ImplUniCharTabData
 116 {
 117     sal_uInt16                      mnUniChar;
 118     sal_uChar                       mnChar;
 119     sal_uChar                       mnChar2;
 120         // to cater for mappings like MS1258 with 1--2 bytes per Unicode char,
 121         // 0 if unused
 122 };
 123
 124 struct ImplByteConvertData
 125 {
 126     const sal_uInt16*               mpToUniTab1;
 127     const sal_uInt16*               mpToUniTab2;
 128     sal_uChar                       mnToUniStart1;
 129     sal_uChar                       mnToUniEnd1;
 130     sal_uChar                       mnToUniStart2;
 131     sal_uChar                       mnToUniEnd2;
 132     const sal_uChar*                mpToCharTab1;
 133     const sal_uChar*                mpToCharTab2;
 134     const ImplUniCharTabData*       mpToCharExTab;
 135     sal_uInt16                      mnToCharStart1;
 136     sal_uInt16                      mnToCharEnd1;
 137     sal_uInt16                      mnToCharStart2;
 138     sal_uInt16                      mnToCharEnd2;
 139     sal_uInt16                      mnToCharExCount;
 140 };
 141
 142 /* ----------------------------------- */
 143 /* - TextConverter - DBCS-Structures - */
 144 /* ----------------------------------- */
 145
 146 struct ImplDBCSEUDCData
 147 {
 148     sal_uChar                       mnLeadStart;
 149     sal_uChar                       mnLeadEnd;
 150     sal_uChar                       mnTrail1Start;
 151     sal_uChar                       mnTrail1End;
 152     sal_uChar                       mnTrail2Start;
 153     sal_uChar                       mnTrail2End;
 154     sal_uChar                       mnTrail3Start;
 155     sal_uChar                       mnTrail3End;
 156     sal_uChar                       mnTrailCount;
 157     sal_uInt16                      mnTrailRangeCount;
 158     sal_uInt16                      mnUniStart;
 159     sal_uInt16                      mnUniEnd;
 160 };
 161
 162 struct ImplDBCSToUniLeadTab
 163 {
 164     sal_uInt16                      mnUniChar;
 165     sal_uInt8                       mnTrailStart;
 166     sal_uInt8                       mnTrailEnd;
 167     const sal_uInt16*               mpToUniTrailTab;
 168 };
 169
 170 struct ImplUniToDBCSHighTab
 171 {
 172     sal_uInt8                       mnLowStart;
 173     sal_uInt8                       mnLowEnd;
 174     const sal_uInt16*               mpToUniTrailTab;
 175 };
 176
 177 struct ImplDBCSConvertData
 178 {
 179     const ImplDBCSToUniLeadTab*     mpToUniLeadTab;
 180     const ImplUniToDBCSHighTab*     mpToDBCSHighTab;
 181     sal_uChar                       mnLeadStart;
 182     sal_uChar                       mnLeadEnd;
 183     sal_uChar                       mnTrailStart;
 184     sal_uChar                       mnTrailEnd;
 185     const ImplDBCSEUDCData*         mpEUDCTab;
 186     sal_uInt16                      mnEUDCCount;
 187 };
 188
 189 /* ---------------------------------- */
 190 /* - TextConverter - EUC-Structures - */
 191 /* ---------------------------------- */
 192
 193 struct ImplEUCJPConvertData
 194 {
 195     const ImplDBCSToUniLeadTab*     mpJIS0208ToUniLeadTab;
 196     const ImplDBCSToUniLeadTab*     mpJIS0212ToUniLeadTab;
 197     const ImplUniToDBCSHighTab*     mpUniToJIS0208HighTab;
 198     const ImplUniToDBCSHighTab*     mpUniToJIS0212HighTab;
 199 };
 200
 201 /* --------------------------------- */
 202 /* - TextConverter - HelpFunctions - */
 203 /* --------------------------------- */
 204
 205 inline sal_Unicode ImplGetUndefinedUnicodeChar(
 206     sal_uChar cChar, sal_uInt32 nFlags)
 207 {
 208     return ((nFlags & RTL_TEXTTOUNICODE_FLAGS_UNDEFINED_MASK)
 209                    == RTL_TEXTTOUNICODE_FLAGS_UNDEFINED_MAPTOPRIVATE) ?
 210                RTL_TEXTCVT_BYTE_PRIVATE_START + cChar :
 211                RTL_TEXTENC_UNICODE_REPLACEMENT_CHARACTER;
 212 }
 213
 214 /* ----------------------------- */
 215 /* - TextConverter - Functions - */
 216 /* ----------------------------- */
 217
 218 sal_Size ImplDBCSToUnicode( const void* pData, void* pContext,
 219                             const char* pSrcBuf, sal_Size nSrcBytes,
 220                             sal_Unicode* pDestBuf, sal_Size nDestChars,
 221                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 222                             sal_Size* pSrcCvtBytes );
 223 sal_Size ImplUnicodeToDBCS( const void* pData, void* pContext,
 224                             const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 225                             char* pDestBuf, sal_Size nDestBytes,
 226                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 227                             sal_Size* pSrcCvtChars );
 228 sal_Size ImplEUCJPToUnicode( const void* pData,
 229                              void* pContext,
 230                              const char* pSrcBuf, sal_Size nSrcBytes,
 231                              sal_Unicode* pDestBuf, sal_Size nDestChars,
 232                              sal_uInt32 nFlags, sal_uInt32* pInfo,
 233                              sal_Size* pSrcCvtBytes );
 234 sal_Size ImplUnicodeToEUCJP( const void* pData,
 235                              void* pContext,
 236                              const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 237                              char* pDestBuf, sal_Size nDestBytes,
 238                              sal_uInt32 nFlags, sal_uInt32* pInfo,
 239                              sal_Size* pSrcCvtChars );
 240 void* ImplUTF7CreateUTF7TextToUnicodeContext();
 241 void ImplUTF7DestroyTextToUnicodeContext( void* pContext );
 242 void ImplUTF7ResetTextToUnicodeContext( void* pContext );
 243 sal_Size ImplUTF7ToUnicode( const void* pData, void* pContext,
 244                             const char* pSrcBuf, sal_Size nSrcBytes,
 245                             sal_Unicode* pDestBuf, sal_Size nDestChars,
 246                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 247                             sal_Size* pSrcCvtBytes );
 248 void* ImplUTF7CreateUnicodeToTextContext();
 249 void ImplUTF7DestroyUnicodeToTextContext( void* pContext );
 250 void ImplUTF7ResetUnicodeToTextContext( void* pContext );
 251 sal_Size ImplUnicodeToUTF7( const void* pData, void* pContext,
 252                             const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 253                             char* pDestBuf, sal_Size nDestBytes,
 254                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 255                             sal_Size* pSrcCvtChars );
 256
 257 #endif
 258
 259 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */