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