cwstring.pp 10 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378
  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. lock_ansi2ucs4 : integer = -1;
  94. lock_ucs42ansi : integer = -1;
  95. lock_ansi2wide : integer = -1;
  96. lock_wide2ansi : integer = -1;
  97. procedure lockiconv(var lockcount: integer);
  98. begin
  99. while interlockedincrement(lockcount) <> 0 do begin
  100. interlockeddecrement(lockcount);
  101. sleep(0);
  102. end;
  103. end;
  104. procedure unlockiconv(var lockcount: integer);
  105. begin
  106. interlockeddecrement(lockcount);
  107. end;
  108. procedure Wide2AnsiMove(source:pwidechar;var dest:ansistring;len:SizeInt);
  109. var
  110. outlength,
  111. outoffset,
  112. srclen,
  113. outleft : size_t;
  114. srcpos : pwidechar;
  115. destpos: pchar;
  116. mynil : pchar;
  117. my0 : size_t;
  118. begin
  119. mynil:=nil;
  120. my0:=0;
  121. { rought estimation }
  122. setlength(dest,len*3);
  123. outlength:=len*3;
  124. srclen:=len*2;
  125. srcpos:=source;
  126. destpos:=pchar(dest);
  127. outleft:=outlength;
  128. lockiconv(lock_wide2ansi);
  129. while iconv(iconv_wide2ansi,@srcpos,@srclen,@destpos,@outleft)=size_t(-1) do
  130. begin
  131. case fpgetCerrno of
  132. ESysEILSEQ:
  133. begin
  134. { skip and set to '?' }
  135. inc(srcpos);
  136. dec(srclen,2);
  137. destpos^:='?';
  138. inc(destpos);
  139. dec(outleft);
  140. { reset }
  141. iconv(iconv_wide2ansi,@mynil,@my0,@mynil,@my0);
  142. end;
  143. ESysE2BIG:
  144. begin
  145. outoffset:=destpos-pchar(dest);
  146. { extend }
  147. setlength(dest,outlength+len*3);
  148. inc(outleft,len*3);
  149. inc(outlength,len*3);
  150. { string could have been moved }
  151. destpos:=pchar(dest)+outoffset;
  152. end;
  153. else
  154. begin
  155. unlockiconv(lock_wide2ansi);
  156. raise EConvertError.Create('iconv error '+IntToStr(fpgetCerrno));
  157. end;
  158. end;
  159. end;
  160. unlockiconv(lock_wide2ansi);
  161. // truncate string
  162. setlength(dest,length(dest)-outleft);
  163. end;
  164. procedure Ansi2WideMove(source:pchar;var dest:widestring;len:SizeInt);
  165. var
  166. outlength,
  167. outoffset,
  168. outleft : size_t;
  169. srcpos,
  170. destpos: pchar;
  171. mynil : pchar;
  172. my0 : size_t;
  173. begin
  174. mynil:=nil;
  175. my0:=0;
  176. // extra space
  177. outlength:=len+1;
  178. setlength(dest,outlength);
  179. outlength:=len+1;
  180. srcpos:=source;
  181. destpos:=pchar(dest);
  182. outleft:=outlength*2;
  183. lockiconv(lock_ansi2wide);
  184. while iconv(iconv_ansi2wide,@srcpos,@len,@destpos,@outleft)=size_t(-1) do
  185. begin
  186. case fpgetCerrno of
  187. ESysEILSEQ:
  188. begin
  189. { skip and set to '?' }
  190. inc(srcpos);
  191. pwidechar(destpos)^:='?';
  192. inc(destpos,2);
  193. dec(outleft,2);
  194. { reset }
  195. iconv(iconv_wide2ansi,@mynil,@my0,@mynil,@my0);
  196. end;
  197. ESysE2BIG:
  198. begin
  199. outoffset:=destpos-pchar(dest);
  200. { extend }
  201. setlength(dest,outlength+len);
  202. inc(outleft,len*2);
  203. inc(outlength,len);
  204. { string could have been moved }
  205. destpos:=pchar(dest)+outoffset;
  206. end;
  207. else
  208. begin
  209. unlockiconv(lock_ansi2wide);
  210. raise EConvertError.Create('iconv error '+IntToStr(fpgetCerrno));
  211. end;
  212. end;
  213. end;
  214. unlockiconv(lock_ansi2wide);
  215. // truncate string
  216. setlength(dest,length(dest)-outleft div 2);
  217. end;
  218. function LowerWideString(const s : WideString) : WideString;
  219. var
  220. i : SizeInt;
  221. begin
  222. SetLength(result,length(s));
  223. for i:=1 to length(s) do
  224. result[i]:=WideChar(towlower(wint_t(s[i])));
  225. end;
  226. function UpperWideString(const s : WideString) : WideString;
  227. var
  228. i : SizeInt;
  229. begin
  230. SetLength(result,length(s));
  231. for i:=1 to length(s) do
  232. result[i]:=WideChar(towupper(wint_t(s[i])));
  233. end;
  234. procedure Ansi2UCS4Move(source:pchar;var dest:UCS4String;len:SizeInt);
  235. var
  236. outlength,
  237. outoffset,
  238. outleft : size_t;
  239. srcpos,
  240. destpos: pchar;
  241. mynil : pchar;
  242. my0 : size_t;
  243. begin
  244. mynil:=nil;
  245. my0:=0;
  246. // extra space
  247. outlength:=len+1;
  248. setlength(dest,outlength);
  249. outlength:=len+1;
  250. srcpos:=source;
  251. destpos:=pchar(dest);
  252. outleft:=outlength*4;
  253. lockiconv(lock_ansi2ucs4);
  254. while iconv(iconv_ansi2ucs4,@srcpos,@len,@destpos,@outleft)=size_t(-1) do
  255. begin
  256. case fpgetCerrno of
  257. ESysE2BIG:
  258. begin
  259. outoffset:=destpos-pchar(dest);
  260. { extend }
  261. setlength(dest,outlength+len);
  262. inc(outleft,len*4);
  263. inc(outlength,len);
  264. { string could have been moved }
  265. destpos:=pchar(dest)+outoffset;
  266. end;
  267. else
  268. begin
  269. unlockiconv(lock_ansi2ucs4);
  270. raise EConvertError.Create('iconv error '+IntToStr(fpgetCerrno));
  271. end;
  272. end;
  273. end;
  274. unlockiconv(lock_ansi2ucs4);
  275. // truncate string
  276. setlength(dest,length(dest)-outleft div 4);
  277. end;
  278. function CompareWideString(const s1, s2 : WideString) : PtrInt;
  279. var
  280. hs1,hs2 : UCS4String;
  281. begin
  282. hs1:=WideStringToUCS4String(s1);
  283. hs2:=WideStringToUCS4String(s2);
  284. result:=wcscoll(pwchar_t(hs1),pwchar_t(hs2));
  285. end;
  286. function CompareTextWideString(const s1, s2 : WideString): PtrInt;
  287. begin
  288. result:=CompareWideString(UpperWideString(s1),UpperWideString(s2));
  289. end;
  290. function StrCompAnsi(s1,s2 : PChar): PtrInt;
  291. begin
  292. result:=strcoll(s1,s2);
  293. end;
  294. Procedure SetCWideStringManager;
  295. Var
  296. CWideStringManager : TWideStringManager;
  297. begin
  298. CWideStringManager:=widestringmanager;
  299. With CWideStringManager do
  300. begin
  301. Wide2AnsiMoveProc:=@Wide2AnsiMove;
  302. Ansi2WideMoveProc:=@Ansi2WideMove;
  303. UpperWideStringProc:=@UpperWideString;
  304. LowerWideStringProc:=@LowerWideString;
  305. CompareWideStringProc:=@CompareWideString;
  306. CompareTextWideStringProc:=@CompareTextWideString;
  307. {
  308. CharLengthPCharProc
  309. UpperAnsiStringProc
  310. LowerAnsiStringProc
  311. CompareStrAnsiStringProc
  312. CompareTextAnsiStringProc
  313. }
  314. StrCompAnsiStringProc:=@StrCompAnsi;
  315. {
  316. StrICompAnsiStringProc
  317. StrLCompAnsiStringProc
  318. StrLICompAnsiStringProc
  319. StrLowerAnsiStringProc
  320. StrUpperAnsiStringProc
  321. }
  322. end;
  323. SetWideStringManager(CWideStringManager);
  324. end;
  325. initialization
  326. SetCWideStringManager;
  327. { init conversion tables }
  328. iconv_wide2ansi:=iconv_open(nl_langinfo(CODESET),unicode_encoding);
  329. iconv_ansi2wide:=iconv_open(unicode_encoding,nl_langinfo(CODESET));
  330. iconv_ucs42ansi:=iconv_open(nl_langinfo(CODESET),'UCS4');
  331. iconv_ansi2ucs4:=iconv_open('UCS4',nl_langinfo(CODESET));
  332. finalization
  333. iconv_close(iconv_ansi2wide);
  334. end.