git.distorted.org.uk Git - u/mdw/putty/blob - unix/uxucs.c

   1 #include <stdio.h>
   2 #include <stdlib.h>
   3 #include <ctype.h>
   4 #include <locale.h>
   5 #include <limits.h>
   6 #include <wchar.h>
   7
   8 #include <time.h>
   9
  10 #include "putty.h"
  11 #include "charset.h"
  12 #include "terminal.h"
  13 #include "misc.h"
  14
  15 /*
  16  * Unix Unicode-handling routines.
  17  */
  18
  19 int is_dbcs_leadbyte(int codepage, char byte)
  20 {
  21     return 0;                          /* we don't do DBCS */
  22 }
  23
  24 int mb_to_wc(int codepage, int flags, char *mbstr, int mblen,
  25              wchar_t *wcstr, int wclen)
  26 {
  27     if (codepage == DEFAULT_CODEPAGE) {
  28         int n = 0;
  29         mbstate_t state;
  30
  31         memset(&state, 0, sizeof state);
  32         setlocale(LC_CTYPE, "");
  33
  34         while (mblen > 0) {
  35             size_t i = mbrtowc(wcstr+n, mbstr, (size_t)mblen, &state);
  36             if (i == (size_t)-1 || i == (size_t)-2)
  37                 break;
  38             n++;
  39             mbstr += i;
  40             mblen -= i;
  41         }
  42
  43         setlocale(LC_CTYPE, "C");
  44
  45         return n;
  46     } else if (codepage == CS_NONE) {
  47         int n = 0;
  48
  49         while (mblen > 0) {
  50             wcstr[n] = 0xD800 | (mbstr[0] & 0xFF);
  51             n++;
  52             mbstr++;
  53             mblen--;
  54         }
  55
  56         return n;
  57     } else
  58         return charset_to_unicode(&mbstr, &mblen, wcstr, wclen, codepage,
  59                                   NULL, NULL, 0);
  60 }
  61
  62 int wc_to_mb(int codepage, int flags, wchar_t *wcstr, int wclen,
  63              char *mbstr, int mblen, char *defchr, int *defused,
  64              struct unicode_data *ucsdata)
  65 {
  66     /* FIXME: we should remove the defused param completely... */
  67     if (defused)
  68         *defused = 0;
  69
  70     if (codepage == DEFAULT_CODEPAGE) {
  71         char output[MB_LEN_MAX];
  72         mbstate_t state;
  73         int n = 0;
  74
  75         memset(&state, 0, sizeof state);
  76         setlocale(LC_CTYPE, "");
  77
  78         while (wclen > 0) {
  79             int i = wcrtomb(output, wcstr[0], &state);
  80             if (i == (size_t)-1 || i > n - mblen)
  81                 break;
  82             memcpy(mbstr+n, output, i);
  83             n += i;
  84             wcstr++;
  85             wclen--;
  86         }
  87
  88         setlocale(LC_CTYPE, "C");
  89
  90         return n;
  91     } else if (codepage == CS_NONE) {
  92         int n = 0;
  93         while (wclen > 0 && n < mblen) {
  94             if (*wcstr >= 0xD800 && *wcstr < 0xD900)
  95                 mbstr[n++] = (*wcstr & 0xFF);
  96             else if (defchr)
  97                 mbstr[n++] = *defchr;
  98             wcstr++;
  99             wclen--;
 100         }
 101         return n;
 102     } else {
 103         return charset_from_unicode(&wcstr, &wclen, mbstr, mblen, codepage,
 104                                     NULL, defchr?defchr:NULL, defchr?1:0);
 105     }
 106 }
 107
 108 /*
 109  * Return value is TRUE if pterm is to run in direct-to-font mode.
 110  */
 111 int init_ucs(struct unicode_data *ucsdata, char *linecharset,
 112              int utf8_override, int font_charset, int vtmode)
 113 {
 114     int i, ret = 0;
 115
 116     /*
 117      * In the platform-independent parts of the code, font_codepage
 118      * is used only for system DBCS support - which we don't
 119      * support at all. So we set this to something which will never
 120      * be used.
 121      */
 122     ucsdata->font_codepage = -1;
 123
 124     /*
 125      * If utf8_override is set and the POSIX locale settings
 126      * dictate a UTF-8 character set, then just go straight for
 127      * UTF-8.
 128      */
 129     ucsdata->line_codepage = CS_NONE;
 130     if (utf8_override) {
 131         const char *s;
 132         if (((s = getenv("LC_ALL"))   && *s) ||
 133             ((s = getenv("LC_CTYPE")) && *s) ||
 134             ((s = getenv("LANG"))     && *s)) {
 135             if (strstr(s, "UTF-8"))
 136                 ucsdata->line_codepage = CS_UTF8;
 137         }
 138     }
 139
 140     /*
 141      * Failing that, line_codepage should be decoded from the
 142      * specification in conf.
 143      */
 144     if (ucsdata->line_codepage == CS_NONE)
 145         ucsdata->line_codepage = decode_codepage(linecharset);
 146
 147     /*
 148      * If line_codepage is _still_ CS_NONE, we assume we're using
 149      * the font's own encoding. This has been passed in to us, so
 150      * we use that. If it's still CS_NONE after _that_ - i.e. the
 151      * font we were given had an incomprehensible charset - then we
 152      * fall back to using the D800 page.
 153      */
 154     if (ucsdata->line_codepage == CS_NONE)
 155         ucsdata->line_codepage = font_charset;
 156
 157     if (ucsdata->line_codepage == CS_NONE)
 158         ret = 1;
 159
 160     /*
 161      * Set up unitab_line, by translating each individual character
 162      * in the line codepage into Unicode.
 163      */
 164     for (i = 0; i < 256; i++) {
 165         char c[1], *p;
 166         wchar_t wc[1];
 167         int len;
 168         c[0] = i;
 169         p = c;
 170         len = 1;
 171         if (ucsdata->line_codepage == CS_NONE)
 172             ucsdata->unitab_line[i] = 0xD800 | i;
 173         else if (1 == charset_to_unicode(&p, &len, wc, 1,
 174                                          ucsdata->line_codepage,
 175                                          NULL, L"", 0))
 176             ucsdata->unitab_line[i] = wc[0];
 177         else
 178             ucsdata->unitab_line[i] = 0xFFFD;
 179     }
 180
 181     /*
 182      * Set up unitab_xterm. This is the same as unitab_line except
 183      * in the line-drawing regions, where it follows the Unicode
 184      * encoding.
 185      *
 186      * (Note that the strange X encoding of line-drawing characters
 187      * in the bottom 32 glyphs of ISO8859-1 fonts is taken care of
 188      * by the font encoding, which will spot such a font and act as
 189      * if it were in a variant encoding of ISO8859-1.)
 190      */
 191     for (i = 0; i < 256; i++) {
 192         static const wchar_t unitab_xterm_std[32] = {
 193             0x2666, 0x2592, 0x2409, 0x240c, 0x240d, 0x240a, 0x00b0, 0x00b1,
 194             0x2424, 0x240b, 0x2518, 0x2510, 0x250c, 0x2514, 0x253c, 0x23ba,
 195             0x23bb, 0x2500, 0x23bc, 0x23bd, 0x251c, 0x2524, 0x2534, 0x252c,
 196             0x2502, 0x2264, 0x2265, 0x03c0, 0x2260, 0x00a3, 0x00b7, 0x0020
 197         };
 198         static const wchar_t unitab_xterm_poorman[32] =
 199             L"*#****o~**+++++-----++++|****L. ";
 200
 201         const wchar_t *ptr;
 202
 203         if (vtmode == VT_POORMAN)
 204             ptr = unitab_xterm_poorman;
 205         else
 206             ptr = unitab_xterm_std;
 207
 208         if (i >= 0x5F && i < 0x7F)
 209             ucsdata->unitab_xterm[i] = ptr[i & 0x1F];
 210         else
 211             ucsdata->unitab_xterm[i] = ucsdata->unitab_line[i];
 212     }
 213
 214     /*
 215      * Set up unitab_scoacs. The SCO Alternate Character Set is
 216      * simply CP437.
 217      */
 218     for (i = 0; i < 256; i++) {
 219         char c[1], *p;
 220         wchar_t wc[1];
 221         int len;
 222         c[0] = i;
 223         p = c;
 224         len = 1;
 225         if (1 == charset_to_unicode(&p, &len, wc, 1, CS_CP437, NULL, L"", 0))
 226             ucsdata->unitab_scoacs[i] = wc[0];
 227         else
 228             ucsdata->unitab_scoacs[i] = 0xFFFD;
 229     }
 230
 231     /*
 232      * Find the control characters in the line codepage. For
 233      * direct-to-font mode using the D800 hack, we assume 00-1F and
 234      * 7F are controls, but allow 80-9F through. (It's as good a
 235      * guess as anything; and my bet is that half the weird fonts
 236      * used in this way will be IBM or MS code pages anyway.)
 237      */
 238     for (i = 0; i < 256; i++) {
 239         int lineval = ucsdata->unitab_line[i];
 240         if (lineval < ' ' || (lineval >= 0x7F && lineval < 0xA0) ||
 241             (lineval >= 0xD800 && lineval < 0xD820) || (lineval == 0xD87F))
 242             ucsdata->unitab_ctrl[i] = i;
 243         else
 244             ucsdata->unitab_ctrl[i] = 0xFF;
 245     }
 246
 247     return ret;
 248 }
 249
 250 const char *cp_name(int codepage)
 251 {
 252     if (codepage == CS_NONE)
 253         return "Use font encoding";
 254     return charset_to_localenc(codepage);
 255 }
 256
 257 const char *cp_enumerate(int index)
 258 {
 259     int charset;
 260     if (index == 0)
 261         return "Use font encoding";
 262     charset = charset_localenc_nth(index-1);
 263     if (charset == CS_NONE)
 264         return NULL;
 265     return charset_to_localenc(charset);
 266 }
 267
 268 int decode_codepage(char *cp_name)
 269 {
 270     if (!*cp_name)
 271         return CS_NONE;                /* use font encoding */
 272     return charset_from_localenc(cp_name);
 273 }