7#include "core/fxcrt/fx_codepage.h"
13#include "build/build_config.h"
14#include "third_party/base/numerics/safe_math.h"
22const uint16_t kFX_MSDOSThaiUnicodes[128] = {
23 0x20AC, 0x0000, 0x0000, 0x0000, 0x0000, 0x2026, 0x0000, 0x0000, 0x0000,
24 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x2018,
25 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x0000, 0x0000, 0x0000,
26 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x00A0, 0x0E01, 0x0E02, 0x0E03,
27 0x0E04, 0x0E05, 0x0E06, 0x0E07, 0x0E08, 0x0E09, 0x0E0A, 0x0E0B, 0x0E0C,
28 0x0E0D, 0x0E0E, 0x0E0F, 0x0E10, 0x0E11, 0x0E12, 0x0E13, 0x0E14, 0x0E15,
29 0x0E16, 0x0E17, 0x0E18, 0x0E19, 0x0E1A, 0x0E1B, 0x0E1C, 0x0E1D, 0x0E1E,
30 0x0E1F, 0x0E20, 0x0E21, 0x0E22, 0x0E23, 0x0E24, 0x0E25, 0x0E26, 0x0E27,
31 0x0E28, 0x0E29, 0x0E2A, 0x0E2B, 0x0E2C, 0x0E2D, 0x0E2E, 0x0E2F, 0x0E30,
32 0x0E31, 0x0E32, 0x0E33, 0x0E34, 0x0E35, 0x0E36, 0x0E37, 0x0E38, 0x0E39,
33 0x0E3A, 0x0000, 0x0000, 0x0000, 0x0000, 0x0E3F, 0x0E40, 0x0E41, 0x0E42,
34 0x0E43, 0x0E44, 0x0E45, 0x0E46, 0x0E47, 0x0E48, 0x0E49, 0x0E4A, 0x0E4B,
35 0x0E4C, 0x0E4D, 0x0E4E, 0x0E4F, 0x0E50, 0x0E51, 0x0E52, 0x0E53, 0x0E54,
36 0x0E55, 0x0E56, 0x0E57, 0x0E58, 0x0E59, 0x0E5A, 0x0E5B, 0x0000, 0x0000,
40const uint16_t kFX_MSWinEasternEuropeanUnicodes[128] = {
41 0x20AC, 0x0000, 0x201A, 0x0000, 0x201E, 0x2026, 0x2020, 0x2021, 0x0000,
42 0x2030, 0x0160, 0x2039, 0x015A, 0x0164, 0x017D, 0x0179, 0x0000, 0x2018,
43 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x0000, 0x2122, 0x0161,
44 0x203A, 0x015B, 0x0165, 0x017E, 0x017A, 0x00A0, 0x02C7, 0x02D8, 0x0141,
45 0x00A4, 0x0104, 0x00A6, 0x00A7, 0x00A8, 0x00A9, 0x015E, 0x00AB, 0x00AC,
46 0x00AD, 0x00AE, 0x017B, 0x00B0, 0x00B1, 0x02DB, 0x0142, 0x00B4, 0x00B5,
47 0x00B6, 0x00B7, 0x00B8, 0x0105, 0x015F, 0x00BB, 0x013D, 0x02DD, 0x013E,
48 0x017C, 0x0154, 0x00C1, 0x00C2, 0x0102, 0x00C4, 0x0139, 0x0106, 0x00C7,
49 0x010C, 0x00C9, 0x0118, 0x00CB, 0x011A, 0x00CD, 0x00CE, 0x010E, 0x0110,
50 0x0143, 0x0147, 0x00D3, 0x00D4, 0x0150, 0x00D6, 0x00D7, 0x0158, 0x016E,
51 0x00DA, 0x0170, 0x00DC, 0x00DD, 0x0162, 0x00DF, 0x0155, 0x00E1, 0x00E2,
52 0x0103, 0x00E4, 0x013A, 0x0107, 0x00E7, 0x010D, 0x00E9, 0x0119, 0x00EB,
53 0x011B, 0x00ED, 0x00EE, 0x010F, 0x0111, 0x0144, 0x0148, 0x00F3, 0x00F4,
54 0x0151, 0x00F6, 0x00F7, 0x0159, 0x016F, 0x00FA, 0x0171, 0x00FC, 0x00FD,
58const uint16_t kFX_MSWinCyrillicUnicodes[128] = {
59 0x0402, 0x0403, 0x201A, 0x0453, 0x201E, 0x2026, 0x2020, 0x2021, 0x20AC,
60 0x2030, 0x0409, 0x2039, 0x040A, 0x040C, 0x040B, 0x040F, 0x0452, 0x2018,
61 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x0000, 0x2122, 0x0459,
62 0x203A, 0x045A, 0x045C, 0x045B, 0x045F, 0x00A0, 0x040E, 0x045E, 0x0408,
63 0x00A4, 0x0490, 0x00A6, 0x00A7, 0x0401, 0x00A9, 0x0404, 0x00AB, 0x00AC,
64 0x00AD, 0x00AE, 0x0407, 0x00B0, 0x00B1, 0x0406, 0x0456, 0x0491, 0x00B5,
65 0x00B6, 0x00B7, 0x0451, 0x2116, 0x0454, 0x00BB, 0x0458, 0x0405, 0x0455,
66 0x0457, 0x0410, 0x0411, 0x0412, 0x0413, 0x0414, 0x0415, 0x0416, 0x0417,
67 0x0418, 0x0419, 0x041A, 0x041B, 0x041C, 0x041D, 0x041E, 0x041F, 0x0420,
68 0x0421, 0x0422, 0x0423, 0x0424, 0x0425, 0x0426, 0x0427, 0x0428, 0x0429,
69 0x042A, 0x042B, 0x042C, 0x042D, 0x042E, 0x042F, 0x0430, 0x0431, 0x0432,
70 0x0433, 0x0434, 0x0435, 0x0436, 0x0437, 0x0438, 0x0439, 0x043A, 0x043B,
71 0x043C, 0x043D, 0x043E, 0x043F, 0x0440, 0x0441, 0x0442, 0x0443, 0x0444,
72 0x0445, 0x0446, 0x0447, 0x0448, 0x0449, 0x044A, 0x044B, 0x044C, 0x044D,
76const uint16_t kFX_MSWinGreekUnicodes[128] = {
77 0x20AC, 0x0000, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021, 0x0000,
78 0x2030, 0x0000, 0x2039, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x2018,
79 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x0000, 0x2122, 0x0000,
80 0x203A, 0x0000, 0x0000, 0x0000, 0x0000, 0x00A0, 0x0385, 0x0386, 0x00A3,
81 0x00A4, 0x00A5, 0x00A6, 0x00A7, 0x00A8, 0x00A9, 0x0000, 0x00AB, 0x00AC,
82 0x00AD, 0x00AE, 0x2015, 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x0384, 0x00B5,
83 0x00B6, 0x00B7, 0x0388, 0x0389, 0x038A, 0x00BB, 0x038C, 0x00BD, 0x038E,
84 0x038F, 0x0390, 0x0391, 0x0392, 0x0393, 0x0394, 0x0395, 0x0396, 0x0397,
85 0x0398, 0x0399, 0x039A, 0x039B, 0x039C, 0x039D, 0x039E, 0x039F, 0x03A0,
86 0x03A1, 0x0000, 0x03A3, 0x03A4, 0x03A5, 0x03A6, 0x03A7, 0x03A8, 0x03A9,
87 0x03AA, 0x03AB, 0x03AC, 0x03AD, 0x03AE, 0x03AF, 0x03B0, 0x03B1, 0x03B2,
88 0x03B3, 0x03B4, 0x03B5, 0x03B6, 0x03B7, 0x03B8, 0x03B9, 0x03BA, 0x03BB,
89 0x03BC, 0x03BD, 0x03BE, 0x03BF, 0x03C0, 0x03C1, 0x03C2, 0x03C3, 0x03C4,
90 0x03C5, 0x03C6, 0x03C7, 0x03C8, 0x03C9, 0x03CA, 0x03CB, 0x03CC, 0x03CD,
94const uint16_t kFX_MSWinTurkishUnicodes[128] = {
95 0x20AC, 0x0000, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021, 0x02C6,
96 0x2030, 0x0160, 0x2039, 0x0152, 0x0000, 0x0000, 0x0000, 0x0000, 0x2018,
97 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x02DC, 0x2122, 0x0161,
98 0x203A, 0x0153, 0x0000, 0x0000, 0x0178, 0x00A0, 0x00A1, 0x00A2, 0x00A3,
99 0x00A4, 0x00A5, 0x00A6, 0x00A7, 0x00A8, 0x00A9, 0x00AA, 0x00AB, 0x00AC,
100 0x00AD, 0x00AE, 0x00AF, 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x00B4, 0x00B5,
101 0x00B6, 0x00B7, 0x00B8, 0x00B9, 0x00BA, 0x00BB, 0x00BC, 0x00BD, 0x00BE,
102 0x00BF, 0x00C0, 0x00C1, 0x00C2, 0x00C3, 0x00C4, 0x00C5, 0x00C6, 0x00C7,
103 0x00C8, 0x00C9, 0x00CA, 0x00CB, 0x00CC, 0x00CD, 0x00CE, 0x00CF, 0x011E,
104 0x00D1, 0x00D2, 0x00D3, 0x00D4, 0x00D5, 0x00D6, 0x00D7, 0x00D8, 0x00D9,
105 0x00DA, 0x00DB, 0x00DC, 0x0130, 0x015E, 0x00DF, 0x00E0, 0x00E1, 0x00E2,
106 0x00E3, 0x00E4, 0x00E5, 0x00E6, 0x00E7, 0x00E8, 0x00E9, 0x00EA, 0x00EB,
107 0x00EC, 0x00ED, 0x00EE, 0x00EF, 0x011F, 0x00F1, 0x00F2, 0x00F3, 0x00F4,
108 0x00F5, 0x00F6, 0x00F7, 0x00F8, 0x00F9, 0x00FA, 0x00FB, 0x00FC, 0x0131,
112const uint16_t kFX_MSWinHebrewUnicodes[128] = {
113 0x20AC, 0x0000, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021, 0x02C6,
114 0x2030, 0x0000, 0x2039, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x2018,
115 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x02DC, 0x2122, 0x0000,
116 0x203A, 0x0000, 0x0000, 0x0000, 0x0000, 0x00A0, 0x00A1, 0x00A2, 0x00A3,
117 0x20AA, 0x00A5, 0x00A6, 0x00A7, 0x00A8, 0x00A9, 0x00D7, 0x00AB, 0x00AC,
118 0x00AD, 0x00AE, 0x00AF, 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x00B4, 0x00B5,
119 0x00B6, 0x00B7, 0x00B8, 0x00B9, 0x00F7, 0x00BB, 0x00BC, 0x00BD, 0x00BE,
120 0x00BF, 0x05B0, 0x05B1, 0x05B2, 0x05B3, 0x05B4, 0x05B5, 0x05B6, 0x05B7,
121 0x05B8, 0x05B9, 0x0000, 0x05BB, 0x05BC, 0x05BD, 0x05BE, 0x05BF, 0x05C0,
122 0x05C1, 0x05C2, 0x05C3, 0x05F0, 0x05F1, 0x05F2, 0x05F3, 0x05F4, 0x0000,
123 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x05D0, 0x05D1, 0x05D2,
124 0x05D3, 0x05D4, 0x05D5, 0x05D6, 0x05D7, 0x05D8, 0x05D9, 0x05DA, 0x05DB,
125 0x05DC, 0x05DD, 0x05DE, 0x05DF, 0x05E0, 0x05E1, 0x05E2, 0x05E3, 0x05E4,
126 0x05E5, 0x05E6, 0x05E7, 0x05E8, 0x05E9, 0x05EA, 0x0000, 0x0000, 0x200E,
130const uint16_t kFX_MSWinArabicUnicodes[128] = {
131 0x20AC, 0x067E, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021, 0x02C6,
132 0x2030, 0x0679, 0x2039, 0x0152, 0x0686, 0x0698, 0x0688, 0x06AF, 0x2018,
133 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x06A9, 0x2122, 0x0691,
134 0x203A, 0x0153, 0x200C, 0x200D, 0x06BA, 0x00A0, 0x060C, 0x00A2, 0x00A3,
135 0x00A4, 0x00A5, 0x00A6, 0x00A7, 0x00A8, 0x00A9, 0x06BE, 0x00AB, 0x00AC,
136 0x00AD, 0x00AE, 0x00AF, 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x00B4, 0x00B5,
137 0x00B6, 0x00B7, 0x00B8, 0x00B9, 0x061B, 0x00BB, 0x00BC, 0x00BD, 0x00BE,
138 0x061F, 0x06C1, 0x0621, 0x0622, 0x0623, 0x0624, 0x0625, 0x0626, 0x0627,
139 0x0628, 0x0629, 0x062A, 0x062B, 0x062C, 0x062D, 0x062E, 0x062F, 0x0630,
140 0x0631, 0x0632, 0x0633, 0x0634, 0x0635, 0x0636, 0x00D7, 0x0637, 0x0638,
141 0x0639, 0x063A, 0x0640, 0x0641, 0x0642, 0x0643, 0x00E0, 0x0644, 0x00E2,
142 0x0645, 0x0646, 0x0647, 0x0648, 0x00E7, 0x00E8, 0x00E9, 0x00EA, 0x00EB,
143 0x0649, 0x064A, 0x00EE, 0x00EF, 0x064B, 0x064C, 0x064D, 0x064E, 0x00F4,
144 0x064F, 0x0650, 0x00F7, 0x0651, 0x00F9, 0x0652, 0x00FB, 0x00FC, 0x200E,
148const uint16_t kFX_MSWinBalticUnicodes[128] = {
149 0x20AC, 0x0000, 0x201A, 0x0000, 0x201E, 0x2026, 0x2020, 0x2021, 0x0000,
150 0x2030, 0x0000, 0x2039, 0x0000, 0x00A8, 0x02C7, 0x00B8, 0x0000, 0x2018,
151 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 0x0000, 0x2122, 0x0000,
152 0x203A, 0x0000, 0x00AF, 0x02DB, 0x0000, 0x00A0, 0x0000, 0x00A2, 0x00A3,
153 0x00A4, 0x0000, 0x00A6, 0x00A7, 0x00D8, 0x00A9, 0x0156, 0x00AB, 0x00AC,
154 0x00AD, 0x00AE, 0x00C6, 0x00B0, 0x00B1, 0x00B2, 0x00B3, 0x00B4, 0x00B5,
155 0x00B6, 0x00B7, 0x00F8, 0x00B9, 0x0157, 0x00BB, 0x00BC, 0x00BD, 0x00BE,
156 0x00E6, 0x0104, 0x012E, 0x0100, 0x0106, 0x00C4, 0x00C5, 0x0118, 0x0112,
157 0x010C, 0x00C9, 0x0179, 0x0116, 0x0122, 0x0136, 0x012A, 0x013B, 0x0160,
158 0x0143, 0x0145, 0x00D3, 0x014C, 0x00D5, 0x00D6, 0x00D7, 0x0172, 0x0141,
159 0x015A, 0x016A, 0x00DC, 0x017B, 0x017D, 0x00DF, 0x0105, 0x012F, 0x0101,
160 0x0107, 0x00E4, 0x00E5, 0x0119, 0x0113, 0x010D, 0x00E9, 0x017A, 0x0117,
161 0x0123, 0x0137, 0x012B, 0x013C, 0x0161, 0x0144, 0x0146, 0x00F3, 0x014D,
162 0x00F5, 0x00F6, 0x00F7, 0x0173, 0x0142, 0x015B, 0x016B, 0x00FC, 0x017C,
166struct FX_CHARSET_MAP {
171const FX_CHARSET_MAP kFXCharset2CodePageTable[] = {
220 return static_cast<FX_CodePage>(GetACP());
227 auto* result =
std::lower_bound(
228 std::begin(kFXCharset2CodePageTable),
std::end(kFXCharset2CodePageTable),
229 charset, [](
const FX_CHARSET_MAP& iter,
const FX_Charset& charset) {
230 return iter.charset < charset;
232 if (result !=
std::end(kFXCharset2CodePageTable) &&
233 result->charset == charset) {
234 return result->codepage;
240 for (
const auto& it : kFXCharset2CodePageTable) {
241 if (it.codepage == codepage)
295 pdfium::span<
char> buf) {
297 int input_len = pdfium::base::checked_cast<
int>(wstr.GetLength());
298 int output_len = pdfium::base::checked_cast<
int>(buf.size());
299 return WideCharToMultiByte(
static_cast<UINT>(codepage), 0,
300 wstr.unterminated_c_str(), input_len, buf.data(),
301 output_len,
nullptr,
nullptr);
304 for (size_t i = 0; i < wstr.GetLength(); i++) {
305 if (wstr[i] < 0x100) {
306 if (len < buf.size())
307 buf[len] =
static_cast<
char>(wstr[i]);
317 pdfium::span<
wchar_t> buf) {
319 const int input_len = pdfium::base::checked_cast<
int>(bstr.GetLength());
320 const int output_len = pdfium::base::checked_cast<
int>(buf.size());
321 return MultiByteToWideChar(
static_cast<UINT>(codepage), 0,
322 bstr.unterminated_c_str(), input_len, buf.data(),
326 for (size_t i = 0; i < bstr.GetLength(); i++) {
327 if (wlen < buf.size())
328 buf[wlen] =
reinterpret_cast<uint8_t>(bstr[i]);
FX_CodePage FX_GetCodePageFromCharset(FX_Charset charset)
size_t FX_WideCharToMultiByte(FX_CodePage codepage, WideStringView wstr, pdfium::span< char > buf)
@ kMAC_ChineseTraditional
size_t FX_MultiByteToWideChar(FX_CodePage codepage, ByteStringView bstr, pdfium::span< wchar_t > buf)
const FX_CharsetUnicodes kFX_CharsetUnicodes[8]
FX_Charset FX_GetCharsetFromCodePage(FX_CodePage codepage)
FX_Charset FX_GetCharsetFromInt(int value)
@ kMAC_ChineseTraditional
bool FX_CharSetIsCJK(FX_Charset uCharset)