cwstring.pp 9.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345
  1. {
  2. This file is part of the Free Pascal run time library.
  3. Copyright (c) 2005 by Florian Klaempfl,
  4. member of the Free Pascal development team.
  5. libc based wide string support
  6. See the file COPYING.FPC, included in this distribution,
  7. for details about the copyright.
  8. This program is distributed in the hope that it will be useful,
  9. but WITHOUT ANY WARRANTY; without even the implied warranty of
  10. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
  11. **********************************************************************}
  12. {$mode objfpc}
  13. unit cwstring;
  14. interface
  15. procedure SetCWidestringManager;
  16. implementation
  17. {$linklib c}
  18. {$ifndef linux} // Linux (and maybe glibc platforms in general), have iconv in glibc.
  19. {$ifndef FreeBSD5}
  20. {$linklib iconv}
  21. {$define useiconv}
  22. {$endif}
  23. {$endif linux}
  24. Uses
  25. BaseUnix,
  26. ctypes,
  27. unix,
  28. unixtype,
  29. sysutils,
  30. initc;
  31. Const
  32. {$ifndef useiconv}
  33. libiconvname='c'; // is in libc under Linux.
  34. {$else}
  35. libiconvname='iconv';
  36. {$endif}
  37. { Case-mapping "arrays" }
  38. var
  39. AnsiUpperChars: AnsiString; // 1..255
  40. AnsiLowerChars: AnsiString; // 1..255
  41. WideUpperChars: WideString; // 1..65535
  42. WideLowerChars: WideString; // 1..65535
  43. { the following declarations are from the libc unit for linux so they
  44. might be very linux centric
  45. maybe this needs to be splitted in an os depend way later }
  46. function towlower(__wc:wint_t):wint_t;cdecl;external libiconvname name 'towlower';
  47. function towupper(__wc:wint_t):wint_t;cdecl;external libiconvname name 'towupper';
  48. function wcscoll (__s1:pwchar_t; __s2:pwchar_t):cint;cdecl;external libiconvname name 'wcscoll';
  49. function strcoll (__s1:pchar; __s2:pchar):cint;cdecl;external libiconvname name 'strcoll';
  50. const
  51. {$ifdef linux}
  52. __LC_CTYPE = 0;
  53. _NL_CTYPE_CLASS = (__LC_CTYPE shl 16);
  54. _NL_CTYPE_CODESET_NAME = (_NL_CTYPE_CLASS)+14;
  55. CODESET = _NL_CTYPE_CODESET_NAME;
  56. {$else linux}
  57. {$ifdef darwin}
  58. CODESET = 0;
  59. {$else darwin}
  60. {$ifdef FreeBSD} // actually FreeBSD5. internationalisation is afaik not default on 4.
  61. CODESET = 0;
  62. {$else freebsd}
  63. {$error lookup the value of CODESET in /usr/include/langinfo.h for your OS }
  64. // and while doing it, check if iconv is in libc, and if the symbols are prefixed with iconv_ or libiconv_
  65. {$endif FreeBSD}
  66. {$endif darwin}
  67. {$endif linux}
  68. { unicode encoding name }
  69. {$ifdef FPC_LITTLE_ENDIAN}
  70. unicode_encoding = 'UNICODELITTLE';
  71. {$else FPC_LITTLE_ENDIAN}
  72. unicode_encoding = 'UNICODEBIG';
  73. {$endif FPC_LITTLE_ENDIAN}
  74. type
  75. piconv_t = ^iconv_t;
  76. iconv_t = pointer;
  77. nl_item = cint;
  78. function nl_langinfo(__item:nl_item):pchar;cdecl;external libiconvname name 'nl_langinfo';
  79. {$ifndef Darwin}
  80. function iconv_open(__tocode:pchar; __fromcode:pchar):iconv_t;cdecl;external libiconvname name 'iconv_open';
  81. function iconv(__cd:iconv_t; __inbuf:ppchar; __inbytesleft:psize_t; __outbuf:ppchar; __outbytesleft:psize_t):size_t;cdecl;external libiconvname name 'iconv';
  82. function iconv_close(__cd:iconv_t):cint;cdecl;external libiconvname name 'iconv_close';
  83. {$else}
  84. function iconv_open(__tocode:pchar; __fromcode:pchar):iconv_t;cdecl;external libiconvname name 'libiconv_open';
  85. function iconv(__cd:iconv_t; __inbuf:ppchar; __inbytesleft:psize_t; __outbuf:ppchar; __outbytesleft:psize_t):size_t;cdecl;external libiconvname name 'libiconv';
  86. function iconv_close(__cd:iconv_t):cint;cdecl;external libiconvname name 'libiconv_close';
  87. {$endif}
  88. var
  89. iconv_ansi2ucs4,
  90. iconv_ucs42ansi,
  91. iconv_ansi2wide,
  92. iconv_wide2ansi : iconv_t;
  93. procedure Wide2AnsiMove(source:pwidechar;var dest:ansistring;len:SizeInt);
  94. var
  95. outlength,
  96. outoffset,
  97. srclen,
  98. outleft : size_t;
  99. srcpos : pwidechar;
  100. destpos: pchar;
  101. mynil : pchar;
  102. my0 : size_t;
  103. begin
  104. mynil:=nil;
  105. my0:=0;
  106. { rought estimation }
  107. setlength(dest,len*3);
  108. outlength:=len*3;
  109. srclen:=len*2;
  110. srcpos:=source;
  111. destpos:=pchar(dest);
  112. outleft:=outlength;
  113. while iconv(iconv_wide2ansi,@srcpos,@srclen,@destpos,@outleft)=size_t(-1) do
  114. begin
  115. case fpgetCerrno of
  116. ESysEILSEQ:
  117. begin
  118. { skip and set to '?' }
  119. inc(srcpos);
  120. dec(srclen,2);
  121. destpos^:='?';
  122. inc(destpos);
  123. dec(outleft);
  124. { reset }
  125. iconv(iconv_wide2ansi,@mynil,@my0,@mynil,@my0);
  126. end;
  127. ESysE2BIG:
  128. begin
  129. outoffset:=destpos-pchar(dest);
  130. { extend }
  131. setlength(dest,outlength+len*3);
  132. inc(outleft,len*3);
  133. inc(outlength,len*3);
  134. { string could have been moved }
  135. destpos:=pchar(dest)+outoffset;
  136. end;
  137. else
  138. raise EConvertError.Create('iconv error '+IntToStr(fpgetCerrno));
  139. end;
  140. end;
  141. // truncate string
  142. setlength(dest,length(dest)-outleft);
  143. end;
  144. procedure Ansi2WideMove(source:pchar;var dest:widestring;len:SizeInt);
  145. var
  146. outlength,
  147. outoffset,
  148. outleft : size_t;
  149. srcpos,
  150. destpos: pchar;
  151. mynil : pchar;
  152. my0 : size_t;
  153. begin
  154. mynil:=nil;
  155. my0:=0;
  156. // extra space
  157. outlength:=len+1;
  158. setlength(dest,outlength);
  159. outlength:=len+1;
  160. srcpos:=source;
  161. destpos:=pchar(dest);
  162. outleft:=outlength*2;
  163. while iconv(iconv_ansi2wide,@srcpos,@len,@destpos,@outleft)=size_t(-1) do
  164. begin
  165. case fpgetCerrno of
  166. ESysEILSEQ:
  167. begin
  168. { skip and set to '?' }
  169. inc(srcpos);
  170. pwidechar(destpos)^:='?';
  171. inc(destpos,2);
  172. dec(outleft,2);
  173. { reset }
  174. iconv(iconv_wide2ansi,@mynil,@my0,@mynil,@my0);
  175. end;
  176. ESysE2BIG:
  177. begin
  178. outoffset:=destpos-pchar(dest);
  179. { extend }
  180. setlength(dest,outlength+len);
  181. inc(outleft,len*2);
  182. inc(outlength,len);
  183. { string could have been moved }
  184. destpos:=pchar(dest)+outoffset;
  185. end;
  186. else
  187. raise EConvertError.Create('iconv error '+IntToStr(fpgetCerrno));
  188. end;
  189. end;
  190. // truncate string
  191. setlength(dest,length(dest)-outleft div 2);
  192. end;
  193. function LowerWideString(const s : WideString) : WideString;
  194. var
  195. i : SizeInt;
  196. begin
  197. SetLength(result,length(s));
  198. for i:=1 to length(s) do
  199. result[i]:=WideChar(towlower(wint_t(s[i])));
  200. end;
  201. function UpperWideString(const s : WideString) : WideString;
  202. var
  203. i : SizeInt;
  204. begin
  205. SetLength(result,length(s));
  206. for i:=1 to length(s) do
  207. result[i]:=WideChar(towupper(wint_t(s[i])));
  208. end;
  209. procedure Ansi2UCS4Move(source:pchar;var dest:UCS4String;len:SizeInt);
  210. var
  211. outlength,
  212. outoffset,
  213. outleft : size_t;
  214. srcpos,
  215. destpos: pchar;
  216. mynil : pchar;
  217. my0 : size_t;
  218. begin
  219. mynil:=nil;
  220. my0:=0;
  221. // extra space
  222. outlength:=len+1;
  223. setlength(dest,outlength);
  224. outlength:=len+1;
  225. srcpos:=source;
  226. destpos:=pchar(dest);
  227. outleft:=outlength*4;
  228. while iconv(iconv_ansi2ucs4,@srcpos,@len,@destpos,@outleft)=size_t(-1) do
  229. begin
  230. case fpgetCerrno of
  231. ESysE2BIG:
  232. begin
  233. outoffset:=destpos-pchar(dest);
  234. { extend }
  235. setlength(dest,outlength+len);
  236. inc(outleft,len*4);
  237. inc(outlength,len);
  238. { string could have been moved }
  239. destpos:=pchar(dest)+outoffset;
  240. end;
  241. else
  242. raise EConvertError.Create('iconv error '+IntToStr(fpgetCerrno));
  243. end;
  244. end;
  245. // truncate string
  246. setlength(dest,length(dest)-outleft div 4);
  247. end;
  248. function CompareWideString(const s1, s2 : WideString) : PtrInt;
  249. var
  250. hs1,hs2 : UCS4String;
  251. begin
  252. hs1:=WideStringToUCS4String(s1);
  253. hs2:=WideStringToUCS4String(s2);
  254. result:=wcscoll(pwchar_t(hs1),pwchar_t(hs2));
  255. end;
  256. function CompareTextWideString(const s1, s2 : WideString): PtrInt;
  257. begin
  258. result:=CompareWideString(UpperWideString(s1),UpperWideString(s2));
  259. end;
  260. function StrCompAnsi(s1,s2 : PChar): PtrInt;
  261. begin
  262. result:=strcoll(s1,s2);
  263. end;
  264. Procedure SetCWideStringManager;
  265. Var
  266. CWideStringManager : TWideStringManager;
  267. begin
  268. CWideStringManager:=widestringmanager;
  269. With CWideStringManager do
  270. begin
  271. Wide2AnsiMoveProc:=@Wide2AnsiMove;
  272. Ansi2WideMoveProc:=@Ansi2WideMove;
  273. UpperWideStringProc:=@UpperWideString;
  274. LowerWideStringProc:=@LowerWideString;
  275. CompareWideStringProc:=@CompareWideString;
  276. CompareTextWideStringProc:=@CompareTextWideString;
  277. {
  278. CharLengthPCharProc
  279. UpperAnsiStringProc
  280. LowerAnsiStringProc
  281. CompareStrAnsiStringProc
  282. CompareTextAnsiStringProc
  283. }
  284. StrCompAnsiStringProc:=@StrCompAnsi;
  285. {
  286. StrICompAnsiStringProc
  287. StrLCompAnsiStringProc
  288. StrLICompAnsiStringProc
  289. StrLowerAnsiStringProc
  290. StrUpperAnsiStringProc
  291. }
  292. end;
  293. SetWideStringManager(CWideStringManager);
  294. end;
  295. initialization
  296. SetCWideStringManager;
  297. { init conversion tables }
  298. iconv_wide2ansi:=iconv_open(nl_langinfo(CODESET),unicode_encoding);
  299. iconv_ansi2wide:=iconv_open(unicode_encoding,nl_langinfo(CODESET));
  300. iconv_ucs42ansi:=iconv_open(nl_langinfo(CODESET),'UCS4');
  301. iconv_ansi2ucs4:=iconv_open('UCS4',nl_langinfo(CODESET));
  302. finalization
  303. iconv_close(iconv_ansi2wide);
  304. end.