sal/textenc/tenchelp.h

   1 /*************************************************************************
   2  *
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * Copyright 2000, 2010 Oracle and/or its affiliates.
   6  *
   7  * OpenOffice.org - a multi-platform office productivity suite
   8  *
   9  * This file is part of OpenOffice.org.
  10  *
  11  * OpenOffice.org is free software: you can redistribute it and/or modify
  12  * it under the terms of the GNU Lesser General Public License version 3
  13  * only, as published by the Free Software Foundation.
  14  *
  15  * OpenOffice.org is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18  * GNU Lesser General Public License version 3 for more details
  19  * (a copy is included in the LICENSE file that accompanied this code).
  20  *
  21  * You should have received a copy of the GNU Lesser General Public License
  22  * version 3 along with OpenOffice.org.  If not, see
  23  * <http://www.openoffice.org/license.html>
  24  * for a copy of the LGPLv3 License.
  25  *
  26  ************************************************************************/
  27
  28 #ifndef INCLUDED_RTL_TEXTENC_TENCHELP_H
  29 #define INCLUDED_RTL_TEXTENC_TENCHELP_H
  30
  31 #include "rtl/tencinfo.h"
  32 #include "rtl/textenc.h"
  33 #include "sal/types.h"
  34
  35 #if defined __cplusplus
  36 extern "C" {
  37 #endif /* __cplusplus */
  38
  39 #define RTL_TEXTCVT_BYTE_PRIVATE_START 0xF100
  40 #define RTL_TEXTCVT_BYTE_PRIVATE_END 0xF1FF
  41
  42 /* ----------------- */
  43 /* - TextConverter - */
  44 /* ----------------- */
  45
  46 typedef void ImplTextConverterData;
  47
  48 typedef
  49 sal_Size (* ImplConvertToUnicodeProc)(ImplTextConverterData const * pData,
  50                                       void * pContext,
  51                                       sal_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)(ImplTextConverterData const * pData,
  61                                    void * pContext,
  62                                    sal_Unicode const * pSrcBuf,
  63                                    sal_Size nSrcChars,
  64                                    sal_Char * pDestBuf,
  65                                    sal_Size nDestBytes,
  66                                    sal_uInt32 nFlags,
  67                                    sal_uInt32 * pInfo,
  68                                    sal_Size * pSrcCvtChars);
  69
  70 typedef void * (* ImplCreateTextContextProc)(void);
  71
  72 typedef void (* ImplDestroyTextContextProc)(void * pContext);
  73
  74 typedef void (* ImplResetTextContextProc)(void * pContext);
  75
  76 typedef void * (* ImplCreateUnicodeContextProc)(void);
  77
  78 typedef void (* ImplDestroyUnicodeContextProc)(void * pContext);
  79
  80 typedef void (* ImplResetUnicodeContextProc)(void * pContext);
  81
  82 typedef struct
  83 {
  84     ImplTextConverterData 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 } ImplTextConverter;
  94
  95 /* ----------------------------- */
  96 /* - TextEncoding - Structures - */
  97 /* ----------------------------- */
  98
  99 typedef struct
 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 } ImplTextEncodingData;
 110
 111 /* ----------------------------------- */
 112 /* - TextConverter - Byte-Structures - */
 113 /* ----------------------------------- */
 114
 115 typedef struct
 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 } ImplUniCharTabData;
 123
 124 typedef struct
 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 } ImplByteConvertData;
 141
 142 /* ----------------------------------- */
 143 /* - TextConverter - DBCS-Structures - */
 144 /* ----------------------------------- */
 145
 146 typedef struct
 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 } ImplDBCSEUDCData;
 161
 162 typedef struct
 163 {
 164     sal_uInt16                      mnUniChar;
 165     sal_uInt8                       mnTrailStart;
 166     sal_uInt8                       mnTrailEnd;
 167     const sal_uInt16*               mpToUniTrailTab;
 168 } ImplDBCSToUniLeadTab;
 169
 170 typedef struct
 171 {
 172     sal_uInt8                       mnLowStart;
 173     sal_uInt8                       mnLowEnd;
 174     const sal_uInt16*               mpToUniTrailTab;
 175 } ImplUniToDBCSHighTab;
 176
 177 typedef struct
 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 } ImplDBCSConvertData;
 188
 189 /* ---------------------------------- */
 190 /* - TextConverter - EUC-Structures - */
 191 /* ---------------------------------- */
 192
 193 typedef struct
 194 {
 195     const ImplDBCSToUniLeadTab*     mpJIS0208ToUniLeadTab;
 196     const ImplDBCSToUniLeadTab*     mpJIS0212ToUniLeadTab;
 197     const ImplUniToDBCSHighTab*     mpUniToJIS0208HighTab;
 198     const ImplUniToDBCSHighTab*     mpUniToJIS0212HighTab;
 199 } ImplEUCJPConvertData;
 200
 201 /* --------------------------------- */
 202 /* - TextConverter - HelpFunctions - */
 203 /* --------------------------------- */
 204
 205 sal_Unicode ImplGetUndefinedUnicodeChar(sal_uChar cChar, sal_uInt32 nFlags);
 206
 207 sal_Bool
 208 ImplHandleUndefinedUnicodeToTextChar(ImplTextConverterData const * pData,
 209                                      sal_Unicode const ** ppSrcBuf,
 210                                      sal_Unicode const * pEndSrcBuf,
 211                                      sal_Char ** ppDestBuf,
 212                                      sal_Char const * pEndDestBuf,
 213                                      sal_uInt32 nFlags,
 214                                      sal_uInt32 * pInfo);
 215     /* sal_True means 'continue,' sal_False means 'break' */
 216
 217 /* ----------------------------- */
 218 /* - TextConverter - Functions - */
 219 /* ----------------------------- */
 220
 221 sal_Size ImplSymbolToUnicode( const ImplTextConverterData* pData, void* pContext,
 222                               const sal_Char* pSrcBuf, sal_Size nSrcBytes,
 223                               sal_Unicode* pDestBuf, sal_Size nDestChars,
 224                               sal_uInt32 nFlags, sal_uInt32* pInfo, sal_Size* pSrcCvtBytes );
 225 sal_Size ImplUnicodeToSymbol( const ImplTextConverterData* pData, void* pContext,
 226                               const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 227                               sal_Char* pDestBuf, sal_Size nDestBytes,
 228                               sal_uInt32 nFlags, sal_uInt32* pInfo, sal_Size* pSrcCvtChars );
 229 sal_Size ImplCharToUnicode( const ImplTextConverterData* pData, void* pContext,
 230                             const sal_Char* pSrcBuf, sal_Size nSrcBytes,
 231                             sal_Unicode* pDestBuf, sal_Size nDestChars,
 232                             sal_uInt32 nFlags, sal_uInt32* pInfo, sal_Size* pSrcCvtBytes );
 233 sal_Size ImplUnicodeToChar( const ImplTextConverterData* pData, void* pContext,
 234                             const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 235                             sal_Char* pDestBuf, sal_Size nDestBytes,
 236                             sal_uInt32 nFlags, sal_uInt32* pInfo, sal_Size* pSrcCvtChars );
 237 sal_Size ImplDBCSToUnicode( const ImplTextConverterData* pData, void* pContext,
 238                             const sal_Char* pSrcBuf, sal_Size nSrcBytes,
 239                             sal_Unicode* pDestBuf, sal_Size nDestChars,
 240                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 241                             sal_Size* pSrcCvtBytes );
 242 sal_Size ImplUnicodeToDBCS( const ImplTextConverterData* pData, void* pContext,
 243                             const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 244                             sal_Char* pDestBuf, sal_Size nDestBytes,
 245                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 246                             sal_Size* pSrcCvtChars );
 247 sal_Size ImplEUCJPToUnicode( const ImplTextConverterData* pData,
 248                              void* pContext,
 249                              const sal_Char* pSrcBuf, sal_Size nSrcBytes,
 250                              sal_Unicode* pDestBuf, sal_Size nDestChars,
 251                              sal_uInt32 nFlags, sal_uInt32* pInfo,
 252                              sal_Size* pSrcCvtBytes );
 253 sal_Size ImplUnicodeToEUCJP( const ImplTextConverterData* pData,
 254                              void* pContext,
 255                              const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 256                              sal_Char* pDestBuf, sal_Size nDestBytes,
 257                              sal_uInt32 nFlags, sal_uInt32* pInfo,
 258                              sal_Size* pSrcCvtChars );
 259 void* ImplUTF7CreateUTF7TextToUnicodeContext( void );
 260 void ImplUTF7DestroyTextToUnicodeContext( void* pContext );
 261 void ImplUTF7ResetTextToUnicodeContext( void* pContext );
 262 sal_Size ImplUTF7ToUnicode( const ImplTextConverterData* pData, void* pContext,
 263                             const sal_Char* pSrcBuf, sal_Size nSrcBytes,
 264                             sal_Unicode* pDestBuf, sal_Size nDestChars,
 265                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 266                             sal_Size* pSrcCvtBytes );
 267 void* ImplUTF7CreateUnicodeToTextContext( void );
 268 void ImplUTF7DestroyUnicodeToTextContext( void* pContext );
 269 void ImplUTF7ResetUnicodeToTextContext( void* pContext );
 270 sal_Size ImplUnicodeToUTF7( const ImplTextConverterData* pData, void* pContext,
 271                             const sal_Unicode* pSrcBuf, sal_Size nSrcChars,
 272                             sal_Char* pDestBuf, sal_Size nDestBytes,
 273                             sal_uInt32 nFlags, sal_uInt32* pInfo,
 274                             sal_Size* pSrcCvtChars );
 275
 276 void * ImplCreateUtf8ToUnicodeContext(void) SAL_THROW_EXTERN_C();
 277 void ImplResetUtf8ToUnicodeContext(void * pContext) SAL_THROW_EXTERN_C();
 278 sal_Size ImplConvertUtf8ToUnicode(ImplTextConverterData const * pData,
 279                                   void * pContext, sal_Char const * pSrcBuf,
 280                                   sal_Size nSrcBytes, sal_Unicode * pDestBuf,
 281                                   sal_Size nDestChars, sal_uInt32 nFlags,
 282                                   sal_uInt32 * pInfo, sal_Size * pSrcCvtBytes)
 283     SAL_THROW_EXTERN_C();
 284 void * ImplCreateUnicodeToUtf8Context(void) SAL_THROW_EXTERN_C();
 285 void ImplResetUnicodeToUtf8Context(void * pContext) SAL_THROW_EXTERN_C();
 286 sal_Size ImplConvertUnicodeToUtf8(ImplTextConverterData const * pData,
 287                                   void * pContext, sal_Unicode const * pSrcBuf,
 288                                   sal_Size nSrcChars, sal_Char * pDestBuf,
 289                                   sal_Size nDestBytes, sal_uInt32 nFlags,
 290                                   sal_uInt32 * pInfo, sal_Size* pSrcCvtChars)
 291     SAL_THROW_EXTERN_C();
 292
 293 #if defined __cplusplus
 294 }
 295 #endif /* __cplusplus */
 296
 297 #endif /* INCLUDED_RTL_TEXTENC_TENCHELP_H */