Index: trunk/src/base/PdfEncoding.cpp =================================================================== --- trunk/src/base/PdfEncoding.cpp (revision 1701) +++ trunk/src/base/PdfEncoding.cpp (working copy) @@ -46,6 +46,8 @@ #include #include #include +#include "PdfArray.h" +#include "doc/PdfDifferenceEncoding.h" namespace PoDoFo { @@ -775,6 +777,36 @@ // ----------------------------------------------------- // // ----------------------------------------------------- +void PdfWinAnsiEncoding::AddToDictionary( PdfDictionary & rDictionary ) const +{ + PdfArray arDifferences; + + for (int i = 0; i < 256; i++) + { + if (PdfWinAnsiEncoding::GetToUnicodeTable()[i] != this->GetToUnicodeTable()[i]) + { + arDifferences.push_back(PdfObject((pdf_int64)i)); + unsigned short shCode = this->GetToUnicodeTable()[i]; +#ifdef PODOFO_IS_LITTLE_ENDIAN + shCode = ((shCode & 0x00FF) << 8) | ((shCode & 0xFF00) >> 8); +#endif + arDifferences.push_back( PdfDifferenceEncoding::UnicodeIDToName(shCode) ); + } + } + + if (!arDifferences.empty()) + { + PdfDictionary dictEncoding; + dictEncoding.AddKey(PdfName("BaseEncoding"), PdfWinAnsiEncoding::GetName()); + dictEncoding.AddKey(PdfName("Differences"), arDifferences); + rDictionary.AddKey(PdfName("Encoding"), dictEncoding); + } + else + { + PdfSimpleEncoding::AddToDictionary(rDictionary); + } +} + const pdf_utf16be* PdfWinAnsiEncoding::GetToUnicodeTable() const { return PdfWinAnsiEncoding::s_cEncoding; @@ -2069,5 +2101,550 @@ 0x27BE, // FE # OPEN-OUTLINED RIGHTWARDS ARROW # a191 0x0000 // FF undefined }; - + +// ----------------------------------------------------- +// PdfWin1250Encoding +// See: http://www.microsoft.com/globaldev/reference/sbcs/1250.mspx +// ----------------------------------------------------- + +// ----------------------------------------------------- +// +// ----------------------------------------------------- +const pdf_utf16be PdfWin1250Encoding::s_cEncoding[256] = { + 0x0000, // NULL + 0x0001, // START OF HEADING + 0x0002, // START OF TEXT + 0x0003, // END OF TEXT + 0x0004, // END OF TRANSMISSION + 0x0005, // ENQUIRY + 0x0006, // ACKNOWLEDGE + 0x0007, // BELL + 0x0008, // BACKSPACE + 0x0009, // HORIZONTAL TABULATION + 0x000A, // LINE FEED + 0x000B, // VERTICAL TABULATION + 0x000C, // FORM FEED + 0x000D, // CARRIAGE RETURN + 0x000E, // SHIFT OUT + 0x000F, // SHIFT IN + 0x0010, // DATA LINK ESCAPE + 0x0011, // DEVICE CONTROL ONE + 0x0012, // DEVICE CONTROL TWO + 0x0013, // DEVICE CONTROL THREE + 0x0014, // DEVICE CONTROL FOUR + 0x0015, // NEGATIVE ACKNOWLEDGE + 0x0016, // SYNCHRONOUS IDLE + 0x0017, // END OF TRANSMISSION BLOCK + 0x0018, // CANCEL + 0x0019, // END OF MEDIUM + 0x001A, // SUBSTITUTE + 0x001B, // ESCAPE + 0x001C, // FILE SEPARATOR + 0x001D, // GROUP SEPARATOR + 0x001E, // RECORD SEPARATOR + 0x001F, // UNIT SEPARATOR + 0x0020, // SPACE + 0x0021, // EXCLAMATION MARK + 0x0022, // QUOTATION MARK + 0x0023, // NUMBER SIGN + 0x0024, // DOLLAR SIGN + 0x0025, // PERCENT SIGN + 0x0026, // AMPERSAND + 0x0027, // APOSTROPHE + 0x0028, // LEFT PARENTHESIS + 0x0029, // RIGHT PARENTHESIS + 0x002A, // ASTERISK + 0x002B, // PLUS SIGN + 0x002C, // COMMA + 0x002D, // HYPHEN-MINUS + 0x002E, // FULL STOP + 0x002F, // SOLIDUS + 0x0030, // DIGIT ZERO + 0x0031, // DIGIT ONE + 0x0032, // DIGIT TWO + 0x0033, // DIGIT THREE + 0x0034, // DIGIT FOUR + 0x0035, // DIGIT FIVE + 0x0036, // DIGIT SIX + 0x0037, // DIGIT SEVEN + 0x0038, // DIGIT EIGHT + 0x0039, // DIGIT NINE + 0x003A, // COLON + 0x003B, // SEMICOLON + 0x003C, // LESS-THAN SIGN + 0x003D, // EQUALS SIGN + 0x003E, // GREATER-THAN SIGN + 0x003F, // QUESTION MARK + 0x0040, // COMMERCIAL AT + 0x0041, // LATIN CAPITAL LETTER A + 0x0042, // LATIN CAPITAL LETTER B + 0x0043, // LATIN CAPITAL LETTER C + 0x0044, // LATIN CAPITAL LETTER D + 0x0045, // LATIN CAPITAL LETTER E + 0x0046, // LATIN CAPITAL LETTER F + 0x0047, // LATIN CAPITAL LETTER G + 0x0048, // LATIN CAPITAL LETTER H + 0x0049, // LATIN CAPITAL LETTER I + 0x004A, // LATIN CAPITAL LETTER J + 0x004B, // LATIN CAPITAL LETTER K + 0x004C, // LATIN CAPITAL LETTER L + 0x004D, // LATIN CAPITAL LETTER M + 0x004E, // LATIN CAPITAL LETTER N + 0x004F, // LATIN CAPITAL LETTER O + 0x0050, // LATIN CAPITAL LETTER P + 0x0051, // LATIN CAPITAL LETTER Q + 0x0052, // LATIN CAPITAL LETTER R + 0x0053, // LATIN CAPITAL LETTER S + 0x0054, // LATIN CAPITAL LETTER T + 0x0055, // LATIN CAPITAL LETTER U + 0x0056, // LATIN CAPITAL LETTER V + 0x0057, // LATIN CAPITAL LETTER W + 0x0058, // LATIN CAPITAL LETTER X + 0x0059, // LATIN CAPITAL LETTER Y + 0x005A, // LATIN CAPITAL LETTER Z + 0x005B, // LEFT SQUARE BRACKET + 0x005C, // REVERSE SOLIDUS + 0x005D, // RIGHT SQUARE BRACKET + 0x005E, // CIRCUMFLEX ACCENT + 0x005F, // LOW LINE + 0x0060, // GRAVE ACCENT + 0x0061, // LATIN SMALL LETTER A + 0x0062, // LATIN SMALL LETTER B + 0x0063, // LATIN SMALL LETTER C + 0x0064, // LATIN SMALL LETTER D + 0x0065, // LATIN SMALL LETTER E + 0x0066, // LATIN SMALL LETTER F + 0x0067, // LATIN SMALL LETTER G + 0x0068, // LATIN SMALL LETTER H + 0x0069, // LATIN SMALL LETTER I + 0x006A, // LATIN SMALL LETTER J + 0x006B, // LATIN SMALL LETTER K + 0x006C, // LATIN SMALL LETTER L + 0x006D, // LATIN SMALL LETTER M + 0x006E, // LATIN SMALL LETTER N + 0x006F, // LATIN SMALL LETTER O + 0x0070, // LATIN SMALL LETTER P + 0x0071, // LATIN SMALL LETTER Q + 0x0072, // LATIN SMALL LETTER R + 0x0073, // LATIN SMALL LETTER S + 0x0074, // LATIN SMALL LETTER T + 0x0075, // LATIN SMALL LETTER U + 0x0076, // LATIN SMALL LETTER V + 0x0077, // LATIN SMALL LETTER W + 0x0078, // LATIN SMALL LETTER X + 0x0079, // LATIN SMALL LETTER Y + 0x007A, // LATIN SMALL LETTER Z + 0x007B, // LEFT CURLY BRACKET + 0x007C, // VERTICAL LINE + 0x007D, // RIGHT CURLY BRACKET + 0x007E, // TILDE + 0x007F, // DELETE + 0x20AC, // EURO SIGN + 0x0000, + 0x201A, // SINGLE LOW-9 QUOTATION MARK + 0x0000, + 0x201E, // DOUBLE LOW-9 QUOTATION MARK + 0x2026, // HORIZONTAL ELLIPSIS + 0x2020, // DAGGER + 0x2021, // DOUBLE DAGGER + 0x0000, + 0x2030, // PER MILLE SIGN + 0x0160, // LATIN CAPITAL LETTER S WITH CARON + 0x2039, // SINGLE LEFT-POINTING ANGLE QUOTATION MARK + 0x015A, // LATIN CAPITAL LETTER S WITH ACUTE + 0x0164, // LATIN CAPITAL LETTER T WITH CARON + 0x017D, // LATIN CAPITAL LETTER Z WITH CARON + 0x0179, // LATIN CAPITAL LETTER Z WITH ACUTE + 0x0000, + 0x2018, // LEFT SINGLE QUOTATION MARK + 0x2019, // RIGHT SINGLE QUOTATION MARK + 0x201C, // LEFT DOUBLE QUOTATION MARK + 0x201D, // RIGHT DOUBLE QUOTATION MARK + 0x2022, // BULLET + 0x2013, // EN DASH + 0x2014, // EM DASH + 0x0000, + 0x2122, // TRADE MARK SIGN + 0x0161, // LATIN SMALL LETTER S WITH CARON + 0x203A, // SINGLE RIGHT-POINTING ANGLE QUOTATION MARK + 0x015B, // LATIN SMALL LETTER S WITH ACUTE + 0x0165, // LATIN SMALL LETTER T WITH CARON + 0x017E, // LATIN SMALL LETTER Z WITH CARON + 0x017A, // LATIN SMALL LETTER Z WITH ACUTE + 0x00A0, // NO-BREAK SPACE + 0x02C7, // CARON + 0x02D8, // BREVE + 0x0141, // LATIN CAPITAL LETTER L WITH STROKE + 0x00A4, // CURRENCY SIGN + 0x0104, // LATIN CAPITAL LETTER A WITH OGONEK + 0x00A6, // BROKEN BAR + 0x00A7, // SECTION SIGN + 0x00A8, // DIAERESIS + 0x00A9, // COPYRIGHT SIGN + 0x015E, // LATIN CAPITAL LETTER S WITH CEDILLA + 0x00AB, // LEFT-POINTING DOUBLE ANGLE QUOTATION MARK + 0x00AC, // NOT SIGN + 0x00AD, // SOFT HYPHEN + 0x00AE, // REGISTERED SIGN + 0x017B, // LATIN CAPITAL LETTER Z WITH DOT ABOVE + 0x00B0, // DEGREE SIGN + 0x00B1, // PLUS-MINUS SIGN + 0x02DB, // OGONEK + 0x0142, // LATIN SMALL LETTER L WITH STROKE + 0x00B4, // ACUTE ACCENT + 0x00B5, // MICRO SIGN + 0x00B6, // PILCROW SIGN + 0x00B7, // MIDDLE DOT + 0x00B8, // CEDILLA + 0x0105, // LATIN SMALL LETTER A WITH OGONEK + 0x015F, // LATIN SMALL LETTER S WITH CEDILLA + 0x00BB, // RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK + 0x013D, // LATIN CAPITAL LETTER L WITH CARON + 0x02DD, // DOUBLE ACUTE ACCENT + 0x013E, // LATIN SMALL LETTER L WITH CARON + 0x017C, // LATIN SMALL LETTER Z WITH DOT ABOVE + 0x0154, // LATIN CAPITAL LETTER R WITH ACUTE + 0x00C1, // LATIN CAPITAL LETTER A WITH ACUTE + 0x00C2, // LATIN CAPITAL LETTER A WITH CIRCUMFLEX + 0x0102, // LATIN CAPITAL LETTER A WITH BREVE + 0x00C4, // LATIN CAPITAL LETTER A WITH DIAERESIS + 0x0139, // LATIN CAPITAL LETTER L WITH ACUTE + 0x0106, // LATIN CAPITAL LETTER C WITH ACUTE + 0x00C7, // LATIN CAPITAL LETTER C WITH CEDILLA + 0x010C, // LATIN CAPITAL LETTER C WITH CARON + 0x00C9, // LATIN CAPITAL LETTER E WITH ACUTE + 0x0118, // LATIN CAPITAL LETTER E WITH OGONEK + 0x00CB, // LATIN CAPITAL LETTER E WITH DIAERESIS + 0x011A, // LATIN CAPITAL LETTER E WITH CARON + 0x00CD, // LATIN CAPITAL LETTER I WITH ACUTE + 0x00CE, // LATIN CAPITAL LETTER I WITH CIRCUMFLEX + 0x010E, // LATIN CAPITAL LETTER D WITH CARON + 0x0110, // LATIN CAPITAL LETTER D WITH STROKE + 0x0143, // LATIN CAPITAL LETTER N WITH ACUTE + 0x0147, // LATIN CAPITAL LETTER N WITH CARON + 0x00D3, // LATIN CAPITAL LETTER O WITH ACUTE + 0x00D4, // LATIN CAPITAL LETTER O WITH CIRCUMFLEX + 0x0150, // LATIN CAPITAL LETTER O WITH DOUBLE ACUTE + 0x00D6, // LATIN CAPITAL LETTER O WITH DIAERESIS + 0x00D7, // MULTIPLICATION SIGN + 0x0158, // LATIN CAPITAL LETTER R WITH CARON + 0x016E, // LATIN CAPITAL LETTER U WITH RING ABOVE + 0x00DA, // LATIN CAPITAL LETTER U WITH ACUTE + 0x0170, // LATIN CAPITAL LETTER U WITH DOUBLE ACUTE + 0x00DC, // LATIN CAPITAL LETTER U WITH DIAERESIS + 0x00DD, // LATIN CAPITAL LETTER Y WITH ACUTE + 0x0162, // LATIN CAPITAL LETTER T WITH CEDILLA + 0x00DF, // LATIN SMALL LETTER SHARP S + 0x0155, // LATIN SMALL LETTER R WITH ACUTE + 0x00E1, // LATIN SMALL LETTER A WITH ACUTE + 0x00E2, // LATIN SMALL LETTER A WITH CIRCUMFLEX + 0x0103, // LATIN SMALL LETTER A WITH BREVE + 0x00E4, // LATIN SMALL LETTER A WITH DIAERESIS + 0x013A, // LATIN SMALL LETTER L WITH ACUTE + 0x0107, // LATIN SMALL LETTER C WITH ACUTE + 0x00E7, // LATIN SMALL LETTER C WITH CEDILLA + 0x010D, // LATIN SMALL LETTER C WITH CARON + 0x00E9, // LATIN SMALL LETTER E WITH ACUTE + 0x0119, // LATIN SMALL LETTER E WITH OGONEK + 0x00EB, // LATIN SMALL LETTER E WITH DIAERESIS + 0x011B, // LATIN SMALL LETTER E WITH CARON + 0x00ED, // LATIN SMALL LETTER I WITH ACUTE + 0x00EE, // LATIN SMALL LETTER I WITH CIRCUMFLEX + 0x010F, // LATIN SMALL LETTER D WITH CARON + 0x0111, // LATIN SMALL LETTER D WITH STROKE + 0x0144, // LATIN SMALL LETTER N WITH ACUTE + 0x0148, // LATIN SMALL LETTER N WITH CARON + 0x00F3, // LATIN SMALL LETTER O WITH ACUTE + 0x00F4, // LATIN SMALL LETTER O WITH CIRCUMFLEX + 0x0151, // LATIN SMALL LETTER O WITH DOUBLE ACUTE + 0x00F6, // LATIN SMALL LETTER O WITH DIAERESIS + 0x00F7, // DIVISION SIGN + 0x0159, // LATIN SMALL LETTER R WITH CARON + 0x016F, // LATIN SMALL LETTER U WITH RING ABOVE + 0x00FA, // LATIN SMALL LETTER U WITH ACUTE + 0x0171, // LATIN SMALL LETTER U WITH DOUBLE ACUTE + 0x00FC, // LATIN SMALL LETTER U WITH DIAERESIS + 0x00FD, // LATIN SMALL LETTER Y WITH ACUTE + 0x0163, // LATIN SMALL LETTER T WITH CEDILLA + 0x02D9 // DOT ABOVE +}; + +const pdf_utf16be* PdfWin1250Encoding::GetToUnicodeTable() const +{ + return PdfWin1250Encoding::s_cEncoding; +} + +// ----------------------------------------------------- +// PdfIso88592Encoding +// See: http://unicode.org/Public/MAPPINGS/ISO8859/8859-2.TXT +// ----------------------------------------------------- + +// ----------------------------------------------------- +// +// ----------------------------------------------------- + +const pdf_utf16be PdfIso88592Encoding::s_cEncoding[256] = { + 0x0000, // NULL + 0x0001, // START OF HEADING + 0x0002, // START OF TEXT + 0x0003, // END OF TEXT + 0x0004, // END OF TRANSMISSION + 0x0005, // ENQUIRY + 0x0006, // ACKNOWLEDGE + 0x0007, // BELL + 0x0008, // BACKSPACE + 0x0009, // HORIZONTAL TABULATION + 0x000A, // LINE FEED + 0x000B, // VERTICAL TABULATION + 0x000C, // FORM FEED + 0x000D, // CARRIAGE RETURN + 0x000E, // SHIFT OUT + 0x000F, // SHIFT IN + 0x0010, // DATA LINK ESCAPE + 0x0011, // DEVICE CONTROL ONE + 0x0012, // DEVICE CONTROL TWO + 0x0013, // DEVICE CONTROL THREE + 0x0014, // DEVICE CONTROL FOUR + 0x0015, // NEGATIVE ACKNOWLEDGE + 0x0016, // SYNCHRONOUS IDLE + 0x0017, // END OF TRANSMISSION BLOCK + 0x0018, // CANCEL + 0x0019, // END OF MEDIUM + 0x001A, // SUBSTITUTE + 0x001B, // ESCAPE + 0x001C, // FILE SEPARATOR + 0x001D, // GROUP SEPARATOR + 0x001E, // RECORD SEPARATOR + 0x001F, // UNIT SEPARATOR + 0x0020, // SPACE + 0x0021, // EXCLAMATION MARK + 0x0022, // QUOTATION MARK + 0x0023, // NUMBER SIGN + 0x0024, // DOLLAR SIGN + 0x0025, // PERCENT SIGN + 0x0026, // AMPERSAND + 0x0027, // APOSTROPHE + 0x0028, // LEFT PARENTHESIS + 0x0029, // RIGHT PARENTHESIS + 0x002A, // ASTERISK + 0x002B, // PLUS SIGN + 0x002C, // COMMA + 0x002D, // HYPHEN-MINUS + 0x002E, // FULL STOP + 0x002F, // SOLIDUS + 0x0030, // DIGIT ZERO + 0x0031, // DIGIT ONE + 0x0032, // DIGIT TWO + 0x0033, // DIGIT THREE + 0x0034, // DIGIT FOUR + 0x0035, // DIGIT FIVE + 0x0036, // DIGIT SIX + 0x0037, // DIGIT SEVEN + 0x0038, // DIGIT EIGHT + 0x0039, // DIGIT NINE + 0x003A, // COLON + 0x003B, // SEMICOLON + 0x003C, // LESS-THAN SIGN + 0x003D, // EQUALS SIGN + 0x003E, // GREATER-THAN SIGN + 0x003F, // QUESTION MARK + 0x0040, // COMMERCIAL AT + 0x0041, // LATIN CAPITAL LETTER A + 0x0042, // LATIN CAPITAL LETTER B + 0x0043, // LATIN CAPITAL LETTER C + 0x0044, // LATIN CAPITAL LETTER D + 0x0045, // LATIN CAPITAL LETTER E + 0x0046, // LATIN CAPITAL LETTER F + 0x0047, // LATIN CAPITAL LETTER G + 0x0048, // LATIN CAPITAL LETTER H + 0x0049, // LATIN CAPITAL LETTER I + 0x004A, // LATIN CAPITAL LETTER J + 0x004B, // LATIN CAPITAL LETTER K + 0x004C, // LATIN CAPITAL LETTER L + 0x004D, // LATIN CAPITAL LETTER M + 0x004E, // LATIN CAPITAL LETTER N + 0x004F, // LATIN CAPITAL LETTER O + 0x0050, // LATIN CAPITAL LETTER P + 0x0051, // LATIN CAPITAL LETTER Q + 0x0052, // LATIN CAPITAL LETTER R + 0x0053, // LATIN CAPITAL LETTER S + 0x0054, // LATIN CAPITAL LETTER T + 0x0055, // LATIN CAPITAL LETTER U + 0x0056, // LATIN CAPITAL LETTER V + 0x0057, // LATIN CAPITAL LETTER W + 0x0058, // LATIN CAPITAL LETTER X + 0x0059, // LATIN CAPITAL LETTER Y + 0x005A, // LATIN CAPITAL LETTER Z + 0x005B, // LEFT SQUARE BRACKET + 0x005C, // REVERSE SOLIDUS + 0x005D, // RIGHT SQUARE BRACKET + 0x005E, // CIRCUMFLEX ACCENT + 0x005F, // LOW LINE + 0x0060, // GRAVE ACCENT + 0x0061, // LATIN SMALL LETTER A + 0x0062, // LATIN SMALL LETTER B + 0x0063, // LATIN SMALL LETTER C + 0x0064, // LATIN SMALL LETTER D + 0x0065, // LATIN SMALL LETTER E + 0x0066, // LATIN SMALL LETTER F + 0x0067, // LATIN SMALL LETTER G + 0x0068, // LATIN SMALL LETTER H + 0x0069, // LATIN SMALL LETTER I + 0x006A, // LATIN SMALL LETTER J + 0x006B, // LATIN SMALL LETTER K + 0x006C, // LATIN SMALL LETTER L + 0x006D, // LATIN SMALL LETTER M + 0x006E, // LATIN SMALL LETTER N + 0x006F, // LATIN SMALL LETTER O + 0x0070, // LATIN SMALL LETTER P + 0x0071, // LATIN SMALL LETTER Q + 0x0072, // LATIN SMALL LETTER R + 0x0073, // LATIN SMALL LETTER S + 0x0074, // LATIN SMALL LETTER T + 0x0075, // LATIN SMALL LETTER U + 0x0076, // LATIN SMALL LETTER V + 0x0077, // LATIN SMALL LETTER W + 0x0078, // LATIN SMALL LETTER X + 0x0079, // LATIN SMALL LETTER Y + 0x007A, // LATIN SMALL LETTER Z + 0x007B, // LEFT CURLY BRACKET + 0x007C, // VERTICAL LINE + 0x007D, // RIGHT CURLY BRACKET + 0x007E, // TILDE + 0x007F, // DELETE + 0x0080, // + 0x0081, // + 0x0082, // + 0x0083, // + 0x0084, // + 0x0085, // + 0x0086, // + 0x0087, // + 0x0088, // + 0x0089, // + 0x008A, // + 0x008B, // + 0x008C, // + 0x008D, // + 0x008E, // + 0x008F, // + 0x0090, // + 0x0091, // + 0x0092, // + 0x0093, // + 0x0094, // + 0x0095, // + 0x0096, // + 0x0097, // + 0x0098, // + 0x0099, // + 0x009A, // + 0x009B, // + 0x009C, // + 0x009D, // + 0x009E, // + 0x009F, // + 0x00A0, // NO-BREAK SPACE + 0x0104, // LATIN CAPITAL LETTER A WITH OGONEK + 0x02D8, // BREVE + 0x0141, // LATIN CAPITAL LETTER L WITH STROKE + 0x00A4, // CURRENCY SIGN + 0x013D, // LATIN CAPITAL LETTER L WITH CARON + 0x015A, // LATIN CAPITAL LETTER S WITH ACUTE + 0x00A7, // SECTION SIGN + 0x00A8, // DIAERESIS + 0x0160, // LATIN CAPITAL LETTER S WITH CARON + 0x015E, // LATIN CAPITAL LETTER S WITH CEDILLA + 0x0164, // LATIN CAPITAL LETTER T WITH CARON + 0x0179, // LATIN CAPITAL LETTER Z WITH ACUTE + 0x00AD, // SOFT HYPHEN + 0x017D, // LATIN CAPITAL LETTER Z WITH CARON + 0x017B, // LATIN CAPITAL LETTER Z WITH DOT ABOVE + 0x00B0, // DEGREE SIGN + 0x0105, // LATIN SMALL LETTER A WITH OGONEK + 0x02DB, // OGONEK + 0x0142, // LATIN SMALL LETTER L WITH STROKE + 0x00B4, // ACUTE ACCENT + 0x013E, // LATIN SMALL LETTER L WITH CARON + 0x015B, // LATIN SMALL LETTER S WITH ACUTE + 0x02C7, // CARON + 0x00B8, // CEDILLA + 0x0161, // LATIN SMALL LETTER S WITH CARON + 0x015F, // LATIN SMALL LETTER S WITH CEDILLA + 0x0165, // LATIN SMALL LETTER T WITH CARON + 0x017A, // LATIN SMALL LETTER Z WITH ACUTE + 0x02DD, // DOUBLE ACUTE ACCENT + 0x017E, // LATIN SMALL LETTER Z WITH CARON + 0x017C, // LATIN SMALL LETTER Z WITH DOT ABOVE + 0x0154, // LATIN CAPITAL LETTER R WITH ACUTE + 0x00C1, // LATIN CAPITAL LETTER A WITH ACUTE + 0x00C2, // LATIN CAPITAL LETTER A WITH CIRCUMFLEX + 0x0102, // LATIN CAPITAL LETTER A WITH BREVE + 0x00C4, // LATIN CAPITAL LETTER A WITH DIAERESIS + 0x0139, // LATIN CAPITAL LETTER L WITH ACUTE + 0x0106, // LATIN CAPITAL LETTER C WITH ACUTE + 0x00C7, // LATIN CAPITAL LETTER C WITH CEDILLA + 0x010C, // LATIN CAPITAL LETTER C WITH CARON + 0x00C9, // LATIN CAPITAL LETTER E WITH ACUTE + 0x0118, // LATIN CAPITAL LETTER E WITH OGONEK + 0x00CB, // LATIN CAPITAL LETTER E WITH DIAERESIS + 0x011A, // LATIN CAPITAL LETTER E WITH CARON + 0x00CD, // LATIN CAPITAL LETTER I WITH ACUTE + 0x00CE, // LATIN CAPITAL LETTER I WITH CIRCUMFLEX + 0x010E, // LATIN CAPITAL LETTER D WITH CARON + 0x0110, // LATIN CAPITAL LETTER D WITH STROKE + 0x0143, // LATIN CAPITAL LETTER N WITH ACUTE + 0x0147, // LATIN CAPITAL LETTER N WITH CARON + 0x00D3, // LATIN CAPITAL LETTER O WITH ACUTE + 0x00D4, // LATIN CAPITAL LETTER O WITH CIRCUMFLEX + 0x0150, // LATIN CAPITAL LETTER O WITH DOUBLE ACUTE + 0x00D6, // LATIN CAPITAL LETTER O WITH DIAERESIS + 0x00D7, // MULTIPLICATION SIGN + 0x0158, // LATIN CAPITAL LETTER R WITH CARON + 0x016E, // LATIN CAPITAL LETTER U WITH RING ABOVE + 0x00DA, // LATIN CAPITAL LETTER U WITH ACUTE + 0x0170, // LATIN CAPITAL LETTER U WITH DOUBLE ACUTE + 0x00DC, // LATIN CAPITAL LETTER U WITH DIAERESIS + 0x00DD, // LATIN CAPITAL LETTER Y WITH ACUTE + 0x0162, // LATIN CAPITAL LETTER T WITH CEDILLA + 0x00DF, // LATIN SMALL LETTER SHARP S + 0x0155, // LATIN SMALL LETTER R WITH ACUTE + 0x00E1, // LATIN SMALL LETTER A WITH ACUTE + 0x00E2, // LATIN SMALL LETTER A WITH CIRCUMFLEX + 0x0103, // LATIN SMALL LETTER A WITH BREVE + 0x00E4, // LATIN SMALL LETTER A WITH DIAERESIS + 0x013A, // LATIN SMALL LETTER L WITH ACUTE + 0x0107, // LATIN SMALL LETTER C WITH ACUTE + 0x00E7, // LATIN SMALL LETTER C WITH CEDILLA + 0x010D, // LATIN SMALL LETTER C WITH CARON + 0x00E9, // LATIN SMALL LETTER E WITH ACUTE + 0x0119, // LATIN SMALL LETTER E WITH OGONEK + 0x00EB, // LATIN SMALL LETTER E WITH DIAERESIS + 0x011B, // LATIN SMALL LETTER E WITH CARON + 0x00ED, // LATIN SMALL LETTER I WITH ACUTE + 0x00EE, // LATIN SMALL LETTER I WITH CIRCUMFLEX + 0x010F, // LATIN SMALL LETTER D WITH CARON + 0x0111, // LATIN SMALL LETTER D WITH STROKE + 0x0144, // LATIN SMALL LETTER N WITH ACUTE + 0x0148, // LATIN SMALL LETTER N WITH CARON + 0x00F3, // LATIN SMALL LETTER O WITH ACUTE + 0x00F4, // LATIN SMALL LETTER O WITH CIRCUMFLEX + 0x0151, // LATIN SMALL LETTER O WITH DOUBLE ACUTE + 0x00F6, // LATIN SMALL LETTER O WITH DIAERESIS + 0x00F7, // DIVISION SIGN + 0x0159, // LATIN SMALL LETTER R WITH CARON + 0x016F, // LATIN SMALL LETTER U WITH RING ABOVE + 0x00FA, // LATIN SMALL LETTER U WITH ACUTE + 0x0171, // LATIN SMALL LETTER U WITH DOUBLE ACUTE + 0x00FC, // LATIN SMALL LETTER U WITH DIAERESIS + 0x00FD, // LATIN SMALL LETTER Y WITH ACUTE + 0x0163, // LATIN SMALL LETTER T WITH CEDILLA + 0x02D9 // DOT ABOVE +}; + +const pdf_utf16be* PdfIso88592Encoding::GetToUnicodeTable() const +{ + return PdfIso88592Encoding::s_cEncoding; +} + }; /* namespace PoDoFo */ Index: trunk/src/base/PdfEncoding.h =================================================================== --- trunk/src/base/PdfEncoding.h (revision 1701) +++ trunk/src/base/PdfEncoding.h (working copy) @@ -517,7 +517,10 @@ * contents on PDF pages. * * It is also called CP-1252 encoding. + * This class may be used as base for derived encodings. * + * \see PdfWin1250Encoding + * * Do not allocate this class yourself, as allocations * might be expensive. Try using PdfFont::WinAnsiEncoding. * @@ -547,6 +550,17 @@ */ virtual const pdf_utf16be* GetToUnicodeTable() const; + /** Add this encoding object to a dictionary + * usually be adding an /Encoding key in font dictionaries. + * + * This method generates array of differences into /Encoding + * dictionary if called from derived class with + * different unicode table. + * + * \param rDictionary add the encoding to this dictionary + */ + virtual void AddToDictionary( PdfDictionary & rDictionary ) const; + private: static const pdf_utf16be s_cEncoding[256]; ///< conversion table from WinAnsiEncoding to UTF16 @@ -726,6 +740,66 @@ }; +/** +* Windows 1250 encoding +*/ +class PODOFO_API PdfWin1250Encoding : public PdfWinAnsiEncoding +{ + public: + + /** Create a new PdfWin1250Encoding + */ + PdfWin1250Encoding() + { + } + + protected: + + /** Gets a table of 256 short values which are the + * big endian unicode code points that are assigned + * to the 256 values of this encoding. + * + * This table is used internally to convert an encoded + * string of this encoding to and from unicode. + * + * \returns an array of 256 big endian unicode code points + */ + virtual const pdf_utf16be* GetToUnicodeTable() const; + + private: + static const pdf_utf16be s_cEncoding[256]; ///< conversion table from ZapfDingbatsEncoding to UTF16 +}; + +/** +* ISO8859-2 encoding +*/ +class PODOFO_API PdfIso88592Encoding : public PdfWinAnsiEncoding +{ + public: + + /** Create a new PdfIso88592Encoding + */ + PdfIso88592Encoding() + { + } + + protected: + + /** Gets a table of 256 short values which are the + * big endian unicode code points that are assigned + * to the 256 values of this encoding. + * + * This table is used internally to convert an encoded + * string of this encoding to and from unicode. + * + * \returns an array of 256 big endian unicode code points + */ + virtual const pdf_utf16be* GetToUnicodeTable() const; + + private: + static const pdf_utf16be s_cEncoding[256]; ///< conversion table from ZapfDingbatsEncoding to UTF16 +}; + }; /* namespace PoDoFo */ Index: trunk/src/doc/PdfFontMetricsBase14.h =================================================================== --- trunk/src/doc/PdfFontMetricsBase14.h (revision 1701) +++ trunk/src/doc/PdfFontMetricsBase14.h (working copy) @@ -235,7 +235,13 @@ inline double GetCapHeight() const; -private: + /** Get a glyph ID by Unicode value. This is needed for generation + * array of widths to /Font element. + * + * \param lUnicode the unicode character value + * \returns a glyph ID + * \see PdfFontType1Base14 + */ long GetGlyphIdUnicode( long lUnicode ) const; private : Index: trunk/src/doc/PdfFontType1Base14.cpp =================================================================== --- trunk/src/doc/PdfFontType1Base14.cpp (revision 1701) +++ trunk/src/doc/PdfFontType1Base14.cpp (working copy) @@ -37,6 +37,8 @@ #include "base/PdfDictionary.h" #include "base/PdfEncoding.h" +#include "PdfFontMetricsBase14.h" +#include "base/PdfArray.h" namespace PoDoFo { @@ -71,7 +73,35 @@ this->GetObject()->GetDictionary().AddKey( PdfName::KeySubtype, PdfName("Type1")); this->GetObject()->GetDictionary().AddKey("BaseFont", PdfName( pMetrics->GetFontname() ) ); - + + PdfFontMetricsBase14 *pBase14Metrics = dynamic_cast(pMetrics); + + if (pBase14Metrics == NULL) + { + PODOFO_RAISE_ERROR( ePdfError_InvalidHandle ); + } + + // Width array for correct character spacing + PdfArray arWidths; + for ( int i = m_pEncoding->GetFirstChar(); i <= m_pEncoding->GetLastChar(); i++ ) + { + // Chars below 32 are invisible- don't solve it + if (i < 32) + { + continue; + } + + unsigned short shCode = m_pEncoding->GetCharCode(i); +#ifdef PODOFO_IS_LITTLE_ENDIAN + shCode = ((shCode & 0x00FF) << 8) | ((shCode & 0xFF00) >> 8); +#endif + + arWidths.push_back(PdfObject( (pdf_int64)pBase14Metrics->GetGlyphWidth(pBase14Metrics->GetGlyphIdUnicode(shCode) ))); + } + this->GetObject()->GetDictionary().AddKey("Widths", arWidths ); + this->GetObject()->GetDictionary().AddKey("FirstChar", PdfVariant( static_cast(32) ) ); + this->GetObject()->GetDictionary().AddKey("LastChar", PdfVariant( static_cast(m_pEncoding->GetLastChar()) ) ); + m_pEncoding->AddToDictionary( this->GetObject()->GetDictionary() ); // Add encoding key // pDescriptor->GetDictionary().AddKey( "FontName", this->GetBaseFont() ); //pDescriptor->GetDictionary().AddKey( "FontWeight", (long)m_pMetrics->Weight() );