Home Explore Help
Register Sign In
baishaojie
/
database_sync
1
0
Fork 0
Files Issues 0 Pull Requests 0 Wiki
Tree: f6861f0127
Branches Tags
database_sync
/
vendor
/
github.com
/
pingcap
/
parser
/
charset
/
encoding_table.go

encoding_table.go 14 KB
History Raw

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260 
  1. // Copyright 2015 PingCAP, Inc.
    2. 

  2. //
    3. 

  3. // Licensed under the Apache License, Version 2.0 (the "License");
    4. 

  4. // you may not use this file except in compliance with the License.
    5. 

  5. // You may obtain a copy of the License at
    6. 

  6. //
    7. 

  7. //     http://www.apache.org/licenses/LICENSE-2.0
    8. 

  8. //
    9. 

  9. // Unless required by applicable law or agreed to in writing, software
    10. 

  10. // distributed under the License is distributed on an "AS IS" BASIS,
    11. 

  11. // See the License for the specific language governing permissions and
    12. 

  12. // limitations under the License.
    13. 

  13. 

  14. package charset
    15. 

  15. 

  16. import (
    17. 

  17. 	"strings"
    18. 

  18. 

  19. 	"golang.org/x/text/encoding"
    20. 

  20. 	"golang.org/x/text/encoding/charmap"
    21. 

  21. 	"golang.org/x/text/encoding/japanese"
    22. 

  22. 	"golang.org/x/text/encoding/korean"
    23. 

  23. 	"golang.org/x/text/encoding/simplifiedchinese"
    24. 

  24. 	"golang.org/x/text/encoding/traditionalchinese"
    25. 

  25. 	"golang.org/x/text/encoding/unicode"
    26. 

  26. )
    27. 

  27. 

  28. // Lookup returns the encoding with the specified label, and its canonical
    29. 

  29. // name. It returns nil and the empty string if label is not one of the
    30. 

  30. // standard encodings for HTML. Matching is case-insensitive and ignores
    31. 

  31. // leading and trailing whitespace.
    32. 

  32. func Lookup(label string) (e encoding.Encoding, name string) {
    33. 

  33. 	label = strings.ToLower(strings.Trim(label, "\t\n\r\f "))
    34. 

  34. 	enc := encodings[label]
    35. 

  35. 	return enc.e, enc.name
    36. 

  36. }
    37. 

  37. 

  38. var encodings = map[string]struct {
    39. 

  39. 	e    encoding.Encoding
    40. 

  40. 	name string
    41. 

  41. }{
    42. 

  42. 	"unicode-1-1-utf-8":   {encoding.Nop, "utf-8"},
    43. 

  43. 	"utf-8":               {encoding.Nop, "utf-8"},
    44. 

  44. 	"utf8":                {encoding.Nop, "utf-8"},
    45. 

  45. 	"utf8mb4":             {encoding.Nop, "utf-8"},
    46. 

  46. 	"binary":              {encoding.Nop, "binary"},
    47. 

  47. 	"866":                 {charmap.CodePage866, "ibm866"},
    48. 

  48. 	"cp866":               {charmap.CodePage866, "ibm866"},
    49. 

  49. 	"csibm866":            {charmap.CodePage866, "ibm866"},
    50. 

  50. 	"ibm866":              {charmap.CodePage866, "ibm866"},
    51. 

  51. 	"csisolatin2":         {charmap.ISO8859_2, "iso-8859-2"},
    52. 

  52. 	"iso-8859-2":          {charmap.ISO8859_2, "iso-8859-2"},
    53. 

  53. 	"iso-ir-101":          {charmap.ISO8859_2, "iso-8859-2"},
    54. 

  54. 	"iso8859-2":           {charmap.ISO8859_2, "iso-8859-2"},
    55. 

  55. 	"iso88592":            {charmap.ISO8859_2, "iso-8859-2"},
    56. 

  56. 	"iso_8859-2":          {charmap.ISO8859_2, "iso-8859-2"},
    57. 

  57. 	"iso_8859-2:1987":     {charmap.ISO8859_2, "iso-8859-2"},
    58. 

  58. 	"l2":                  {charmap.ISO8859_2, "iso-8859-2"},
    59. 

  59. 	"latin2":              {charmap.ISO8859_2, "iso-8859-2"},
    60. 

  60. 	"csisolatin3":         {charmap.ISO8859_3, "iso-8859-3"},
    61. 

  61. 	"iso-8859-3":          {charmap.ISO8859_3, "iso-8859-3"},
    62. 

  62. 	"iso-ir-109":          {charmap.ISO8859_3, "iso-8859-3"},
    63. 

  63. 	"iso8859-3":           {charmap.ISO8859_3, "iso-8859-3"},
    64. 

  64. 	"iso88593":            {charmap.ISO8859_3, "iso-8859-3"},
    65. 

  65. 	"iso_8859-3":          {charmap.ISO8859_3, "iso-8859-3"},
    66. 

  66. 	"iso_8859-3:1988":     {charmap.ISO8859_3, "iso-8859-3"},
    67. 

  67. 	"l3":                  {charmap.ISO8859_3, "iso-8859-3"},
    68. 

  68. 	"latin3":              {charmap.ISO8859_3, "iso-8859-3"},
    69. 

  69. 	"csisolatin4":         {charmap.ISO8859_4, "iso-8859-4"},
    70. 

  70. 	"iso-8859-4":          {charmap.ISO8859_4, "iso-8859-4"},
    71. 

  71. 	"iso-ir-110":          {charmap.ISO8859_4, "iso-8859-4"},
    72. 

  72. 	"iso8859-4":           {charmap.ISO8859_4, "iso-8859-4"},
    73. 

  73. 	"iso88594":            {charmap.ISO8859_4, "iso-8859-4"},
    74. 

  74. 	"iso_8859-4":          {charmap.ISO8859_4, "iso-8859-4"},
    75. 

  75. 	"iso_8859-4:1988":     {charmap.ISO8859_4, "iso-8859-4"},
    76. 

  76. 	"l4":                  {charmap.ISO8859_4, "iso-8859-4"},
    77. 

  77. 	"latin4":              {charmap.ISO8859_4, "iso-8859-4"},
    78. 

  78. 	"csisolatincyrillic":  {charmap.ISO8859_5, "iso-8859-5"},
    79. 

  79. 	"cyrillic":            {charmap.ISO8859_5, "iso-8859-5"},
    80. 

  80. 	"iso-8859-5":          {charmap.ISO8859_5, "iso-8859-5"},
    81. 

  81. 	"iso-ir-144":          {charmap.ISO8859_5, "iso-8859-5"},
    82. 

  82. 	"iso8859-5":           {charmap.ISO8859_5, "iso-8859-5"},
    83. 

  83. 	"iso88595":            {charmap.ISO8859_5, "iso-8859-5"},
    84. 

  84. 	"iso_8859-5":          {charmap.ISO8859_5, "iso-8859-5"},
    85. 

  85. 	"iso_8859-5:1988":     {charmap.ISO8859_5, "iso-8859-5"},
    86. 

  86. 	"arabic":              {charmap.ISO8859_6, "iso-8859-6"},
    87. 

  87. 	"asmo-708":            {charmap.ISO8859_6, "iso-8859-6"},
    88. 

  88. 	"csiso88596e":         {charmap.ISO8859_6, "iso-8859-6"},
    89. 

  89. 	"csiso88596i":         {charmap.ISO8859_6, "iso-8859-6"},
    90. 

  90. 	"csisolatinarabic":    {charmap.ISO8859_6, "iso-8859-6"},
    91. 

  91. 	"ecma-114":            {charmap.ISO8859_6, "iso-8859-6"},
    92. 

  92. 	"iso-8859-6":          {charmap.ISO8859_6, "iso-8859-6"},
    93. 

  93. 	"iso-8859-6-e":        {charmap.ISO8859_6, "iso-8859-6"},
    94. 

  94. 	"iso-8859-6-i":        {charmap.ISO8859_6, "iso-8859-6"},
    95. 

  95. 	"iso-ir-127":          {charmap.ISO8859_6, "iso-8859-6"},
    96. 

  96. 	"iso8859-6":           {charmap.ISO8859_6, "iso-8859-6"},
    97. 

  97. 	"iso88596":            {charmap.ISO8859_6, "iso-8859-6"},
    98. 

  98. 	"iso_8859-6":          {charmap.ISO8859_6, "iso-8859-6"},
    99. 

  99. 	"iso_8859-6:1987":     {charmap.ISO8859_6, "iso-8859-6"},
    100. 

  100. 	"csisolatingreek":     {charmap.ISO8859_7, "iso-8859-7"},
    101. 

  101. 	"ecma-118":            {charmap.ISO8859_7, "iso-8859-7"},
    102. 

  102. 	"elot_928":            {charmap.ISO8859_7, "iso-8859-7"},
    103. 

  103. 	"greek":               {charmap.ISO8859_7, "iso-8859-7"},
    104. 

  104. 	"greek8":              {charmap.ISO8859_7, "iso-8859-7"},
    105. 

  105. 	"iso-8859-7":          {charmap.ISO8859_7, "iso-8859-7"},
    106. 

  106. 	"iso-ir-126":          {charmap.ISO8859_7, "iso-8859-7"},
    107. 

  107. 	"iso8859-7":           {charmap.ISO8859_7, "iso-8859-7"},
    108. 

  108. 	"iso88597":            {charmap.ISO8859_7, "iso-8859-7"},
    109. 

  109. 	"iso_8859-7":          {charmap.ISO8859_7, "iso-8859-7"},
    110. 

  110. 	"iso_8859-7:1987":     {charmap.ISO8859_7, "iso-8859-7"},
    111. 

  111. 	"sun_eu_greek":        {charmap.ISO8859_7, "iso-8859-7"},
    112. 

  112. 	"csiso88598e":         {charmap.ISO8859_8, "iso-8859-8"},
    113. 

  113. 	"csisolatinhebrew":    {charmap.ISO8859_8, "iso-8859-8"},
    114. 

  114. 	"hebrew":              {charmap.ISO8859_8, "iso-8859-8"},
    115. 

  115. 	"iso-8859-8":          {charmap.ISO8859_8, "iso-8859-8"},
    116. 

  116. 	"iso-8859-8-e":        {charmap.ISO8859_8, "iso-8859-8"},
    117. 

  117. 	"iso-ir-138":          {charmap.ISO8859_8, "iso-8859-8"},
    118. 

  118. 	"iso8859-8":           {charmap.ISO8859_8, "iso-8859-8"},
    119. 

  119. 	"iso88598":            {charmap.ISO8859_8, "iso-8859-8"},
    120. 

  120. 	"iso_8859-8":          {charmap.ISO8859_8, "iso-8859-8"},
    121. 

  121. 	"iso_8859-8:1988":     {charmap.ISO8859_8, "iso-8859-8"},
    122. 

  122. 	"visual":              {charmap.ISO8859_8, "iso-8859-8"},
    123. 

  123. 	"csiso88598i":         {charmap.ISO8859_8, "iso-8859-8-i"},
    124. 

  124. 	"iso-8859-8-i":        {charmap.ISO8859_8, "iso-8859-8-i"},
    125. 

  125. 	"logical":             {charmap.ISO8859_8, "iso-8859-8-i"},
    126. 

  126. 	"csisolatin6":         {charmap.ISO8859_10, "iso-8859-10"},
    127. 

  127. 	"iso-8859-10":         {charmap.ISO8859_10, "iso-8859-10"},
    128. 

  128. 	"iso-ir-157":          {charmap.ISO8859_10, "iso-8859-10"},
    129. 

  129. 	"iso8859-10":          {charmap.ISO8859_10, "iso-8859-10"},
    130. 

  130. 	"iso885910":           {charmap.ISO8859_10, "iso-8859-10"},
    131. 

  131. 	"l6":                  {charmap.ISO8859_10, "iso-8859-10"},
    132. 

  132. 	"latin6":              {charmap.ISO8859_10, "iso-8859-10"},
    133. 

  133. 	"iso-8859-13":         {charmap.ISO8859_13, "iso-8859-13"},
    134. 

  134. 	"iso8859-13":          {charmap.ISO8859_13, "iso-8859-13"},
    135. 

  135. 	"iso885913":           {charmap.ISO8859_13, "iso-8859-13"},
    136. 

  136. 	"iso-8859-14":         {charmap.ISO8859_14, "iso-8859-14"},
    137. 

  137. 	"iso8859-14":          {charmap.ISO8859_14, "iso-8859-14"},
    138. 

  138. 	"iso885914":           {charmap.ISO8859_14, "iso-8859-14"},
    139. 

  139. 	"csisolatin9":         {charmap.ISO8859_15, "iso-8859-15"},
    140. 

  140. 	"iso-8859-15":         {charmap.ISO8859_15, "iso-8859-15"},
    141. 

  141. 	"iso8859-15":          {charmap.ISO8859_15, "iso-8859-15"},
    142. 

  142. 	"iso885915":           {charmap.ISO8859_15, "iso-8859-15"},
    143. 

  143. 	"iso_8859-15":         {charmap.ISO8859_15, "iso-8859-15"},
    144. 

  144. 	"l9":                  {charmap.ISO8859_15, "iso-8859-15"},
    145. 

  145. 	"iso-8859-16":         {charmap.ISO8859_16, "iso-8859-16"},
    146. 

  146. 	"cskoi8r":             {charmap.KOI8R, "koi8-r"},
    147. 

  147. 	"koi":                 {charmap.KOI8R, "koi8-r"},
    148. 

  148. 	"koi8":                {charmap.KOI8R, "koi8-r"},
    149. 

  149. 	"koi8-r":              {charmap.KOI8R, "koi8-r"},
    150. 

  150. 	"koi8_r":              {charmap.KOI8R, "koi8-r"},
    151. 

  151. 	"koi8-u":              {charmap.KOI8U, "koi8-u"},
    152. 

  152. 	"csmacintosh":         {charmap.Macintosh, "macintosh"},
    153. 

  153. 	"mac":                 {charmap.Macintosh, "macintosh"},
    154. 

  154. 	"macintosh":           {charmap.Macintosh, "macintosh"},
    155. 

  155. 	"x-mac-roman":         {charmap.Macintosh, "macintosh"},
    156. 

  156. 	"dos-874":             {charmap.Windows874, "windows-874"},
    157. 

  157. 	"iso-8859-11":         {charmap.Windows874, "windows-874"},
    158. 

  158. 	"iso8859-11":          {charmap.Windows874, "windows-874"},
    159. 

  159. 	"iso885911":           {charmap.Windows874, "windows-874"},
    160. 

  160. 	"tis-620":             {charmap.Windows874, "windows-874"},
    161. 

  161. 	"windows-874":         {charmap.Windows874, "windows-874"},
    162. 

  162. 	"cp1250":              {charmap.Windows1250, "windows-1250"},
    163. 

  163. 	"windows-1250":        {charmap.Windows1250, "windows-1250"},
    164. 

  164. 	"x-cp1250":            {charmap.Windows1250, "windows-1250"},
    165. 

  165. 	"cp1251":              {charmap.Windows1251, "windows-1251"},
    166. 

  166. 	"windows-1251":        {charmap.Windows1251, "windows-1251"},
    167. 

  167. 	"x-cp1251":            {charmap.Windows1251, "windows-1251"},
    168. 

  168. 	"ansi_x3.4-1968":      {charmap.Windows1252, "windows-1252"},
    169. 

  169. 	"ascii":               {charmap.Windows1252, "windows-1252"},
    170. 

  170. 	"cp1252":              {charmap.Windows1252, "windows-1252"},
    171. 

  171. 	"cp819":               {charmap.Windows1252, "windows-1252"},
    172. 

  172. 	"csisolatin1":         {charmap.Windows1252, "windows-1252"},
    173. 

  173. 	"ibm819":              {charmap.Windows1252, "windows-1252"},
    174. 

  174. 	"iso-8859-1":          {charmap.Windows1252, "windows-1252"},
    175. 

  175. 	"iso-ir-100":          {charmap.Windows1252, "windows-1252"},
    176. 

  176. 	"iso8859-1":           {charmap.Windows1252, "windows-1252"},
    177. 

  177. 	"iso88591":            {charmap.Windows1252, "windows-1252"},
    178. 

  178. 	"iso_8859-1":          {charmap.Windows1252, "windows-1252"},
    179. 

  179. 	"iso_8859-1:1987":     {charmap.Windows1252, "windows-1252"},
    180. 

  180. 	"l1":                  {charmap.Windows1252, "windows-1252"},
    181. 

  181. 	"latin1":              {charmap.Windows1252, "windows-1252"},
    182. 

  182. 	"us-ascii":            {charmap.Windows1252, "windows-1252"},
    183. 

  183. 	"windows-1252":        {charmap.Windows1252, "windows-1252"},
    184. 

  184. 	"x-cp1252":            {charmap.Windows1252, "windows-1252"},
    185. 

  185. 	"cp1253":              {charmap.Windows1253, "windows-1253"},
    186. 

  186. 	"windows-1253":        {charmap.Windows1253, "windows-1253"},
    187. 

  187. 	"x-cp1253":            {charmap.Windows1253, "windows-1253"},
    188. 

  188. 	"cp1254":              {charmap.Windows1254, "windows-1254"},
    189. 

  189. 	"csisolatin5":         {charmap.Windows1254, "windows-1254"},
    190. 

  190. 	"iso-8859-9":          {charmap.Windows1254, "windows-1254"},
    191. 

  191. 	"iso-ir-148":          {charmap.Windows1254, "windows-1254"},
    192. 

  192. 	"iso8859-9":           {charmap.Windows1254, "windows-1254"},
    193. 

  193. 	"iso88599":            {charmap.Windows1254, "windows-1254"},
    194. 

  194. 	"iso_8859-9":          {charmap.Windows1254, "windows-1254"},
    195. 

  195. 	"iso_8859-9:1989":     {charmap.Windows1254, "windows-1254"},
    196. 

  196. 	"l5":                  {charmap.Windows1254, "windows-1254"},
    197. 

  197. 	"latin5":              {charmap.Windows1254, "windows-1254"},
    198. 

  198. 	"windows-1254":        {charmap.Windows1254, "windows-1254"},
    199. 

  199. 	"x-cp1254":            {charmap.Windows1254, "windows-1254"},
    200. 

  200. 	"cp1255":              {charmap.Windows1255, "windows-1255"},
    201. 

  201. 	"windows-1255":        {charmap.Windows1255, "windows-1255"},
    202. 

  202. 	"x-cp1255":            {charmap.Windows1255, "windows-1255"},
    203. 

  203. 	"cp1256":              {charmap.Windows1256, "windows-1256"},
    204. 

  204. 	"windows-1256":        {charmap.Windows1256, "windows-1256"},
    205. 

  205. 	"x-cp1256":            {charmap.Windows1256, "windows-1256"},
    206. 

  206. 	"cp1257":              {charmap.Windows1257, "windows-1257"},
    207. 

  207. 	"windows-1257":        {charmap.Windows1257, "windows-1257"},
    208. 

  208. 	"x-cp1257":            {charmap.Windows1257, "windows-1257"},
    209. 

  209. 	"cp1258":              {charmap.Windows1258, "windows-1258"},
    210. 

  210. 	"windows-1258":        {charmap.Windows1258, "windows-1258"},
    211. 

  211. 	"x-cp1258":            {charmap.Windows1258, "windows-1258"},
    212. 

  212. 	"x-mac-cyrillic":      {charmap.MacintoshCyrillic, "x-mac-cyrillic"},
    213. 

  213. 	"x-mac-ukrainian":     {charmap.MacintoshCyrillic, "x-mac-cyrillic"},
    214. 

  214. 	"chinese":             {simplifiedchinese.GBK, "gbk"},
    215. 

  215. 	"csgb2312":            {simplifiedchinese.GBK, "gbk"},
    216. 

  216. 	"csiso58gb231280":     {simplifiedchinese.GBK, "gbk"},
    217. 

  217. 	"gb2312":              {simplifiedchinese.GBK, "gbk"},
    218. 

  218. 	"gb_2312":             {simplifiedchinese.GBK, "gbk"},
    219. 

  219. 	"gb_2312-80":          {simplifiedchinese.GBK, "gbk"},
    220. 

  220. 	"gbk":                 {simplifiedchinese.GBK, "gbk"},
    221. 

  221. 	"iso-ir-58":           {simplifiedchinese.GBK, "gbk"},
    222. 

  222. 	"x-gbk":               {simplifiedchinese.GBK, "gbk"},
    223. 

  223. 	"gb18030":             {simplifiedchinese.GB18030, "gb18030"},
    224. 

  224. 	"hz-gb-2312":          {simplifiedchinese.HZGB2312, "hz-gb-2312"},
    225. 

  225. 	"big5":                {traditionalchinese.Big5, "big5"},
    226. 

  226. 	"big5-hkscs":          {traditionalchinese.Big5, "big5"},
    227. 

  227. 	"cn-big5":             {traditionalchinese.Big5, "big5"},
    228. 

  228. 	"csbig5":              {traditionalchinese.Big5, "big5"},
    229. 

  229. 	"x-x-big5":            {traditionalchinese.Big5, "big5"},
    230. 

  230. 	"cseucpkdfmtjapanese": {japanese.EUCJP, "euc-jp"},
    231. 

  231. 	"euc-jp":              {japanese.EUCJP, "euc-jp"},
    232. 

  232. 	"x-euc-jp":            {japanese.EUCJP, "euc-jp"},
    233. 

  233. 	"csiso2022jp":         {japanese.ISO2022JP, "iso-2022-jp"},
    234. 

  234. 	"iso-2022-jp":         {japanese.ISO2022JP, "iso-2022-jp"},
    235. 

  235. 	"csshiftjis":          {japanese.ShiftJIS, "shift_jis"},
    236. 

  236. 	"ms_kanji":            {japanese.ShiftJIS, "shift_jis"},
    237. 

  237. 	"shift-jis":           {japanese.ShiftJIS, "shift_jis"},
    238. 

  238. 	"shift_jis":           {japanese.ShiftJIS, "shift_jis"},
    239. 

  239. 	"sjis":                {japanese.ShiftJIS, "shift_jis"},
    240. 

  240. 	"windows-31j":         {japanese.ShiftJIS, "shift_jis"},
    241. 

  241. 	"x-sjis":              {japanese.ShiftJIS, "shift_jis"},
    242. 

  242. 	"cseuckr":             {korean.EUCKR, "euc-kr"},
    243. 

  243. 	"csksc56011987":       {korean.EUCKR, "euc-kr"},
    244. 

  244. 	"euc-kr":              {korean.EUCKR, "euc-kr"},
    245. 

  245. 	"iso-ir-149":          {korean.EUCKR, "euc-kr"},
    246. 

  246. 	"korean":              {korean.EUCKR, "euc-kr"},
    247. 

  247. 	"ks_c_5601-1987":      {korean.EUCKR, "euc-kr"},
    248. 

  248. 	"ks_c_5601-1989":      {korean.EUCKR, "euc-kr"},
    249. 

  249. 	"ksc5601":             {korean.EUCKR, "euc-kr"},
    250. 

  250. 	"ksc_5601":            {korean.EUCKR, "euc-kr"},
    251. 

  251. 	"windows-949":         {korean.EUCKR, "euc-kr"},
    252. 

  252. 	"csiso2022kr":         {encoding.Replacement, "replacement"},
    253. 

  253. 	"iso-2022-kr":         {encoding.Replacement, "replacement"},
    254. 

  254. 	"iso-2022-cn":         {encoding.Replacement, "replacement"},
    255. 

  255. 	"iso-2022-cn-ext":     {encoding.Replacement, "replacement"},
    256. 

  256. 	"utf-16be":            {unicode.UTF16(unicode.BigEndian, unicode.IgnoreBOM), "utf-16be"},
    257. 

  257. 	"utf-16":              {unicode.UTF16(unicode.LittleEndian, unicode.IgnoreBOM), "utf-16le"},
    258. 

  258. 	"utf-16le":            {unicode.UTF16(unicode.LittleEndian, unicode.IgnoreBOM), "utf-16le"},
    259. 

  259. 	"x-user-defined":      {charmap.XUserDefined, "x-user-defined"},
    260. 

  260. }
    261.