| 1 | /* |
| 2 | * This file is part of DisOrder. |
| 3 | * Copyright (C) 2004, 2005 Richard Kettlewell |
| 4 | * |
| 5 | * This program is free software; you can redistribute it and/or modify |
| 6 | * it under the terms of the GNU General Public License as published by |
| 7 | * the Free Software Foundation; either version 2 of the License, or |
| 8 | * (at your option) any later version. |
| 9 | * |
| 10 | * This program is distributed in the hope that it will be useful, but |
| 11 | * WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 13 | * General Public License for more details. |
| 14 | * |
| 15 | * You should have received a copy of the GNU General Public License |
| 16 | * along with this program; if not, write to the Free Software |
| 17 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 |
| 18 | * USA |
| 19 | */ |
| 20 | /** @file lib/charset.c @brief Character set conversion */ |
| 21 | |
| 22 | #include <config.h> |
| 23 | #include "types.h" |
| 24 | |
| 25 | #include <iconv.h> |
| 26 | #include <string.h> |
| 27 | #include <errno.h> |
| 28 | #include <langinfo.h> |
| 29 | |
| 30 | #include "mem.h" |
| 31 | #include "log.h" |
| 32 | #include "charset.h" |
| 33 | #include "configuration.h" |
| 34 | #include "utf8.h" |
| 35 | #include "vector.h" |
| 36 | #include "unidata.h" |
| 37 | |
| 38 | /** @brief Low-level converstion routine |
| 39 | * @param from Source encoding |
| 40 | * @param to Destination encoding |
| 41 | * @param ptr First byte to convert |
| 42 | * @param n Number of bytes to convert |
| 43 | * @return Converted text, 0-terminated; or NULL on error. |
| 44 | */ |
| 45 | static void *convert(const char *from, const char *to, |
| 46 | const void *ptr, size_t n) { |
| 47 | iconv_t i; |
| 48 | size_t len; |
| 49 | char *buf = 0, *s, *d; |
| 50 | size_t bufsize = 0, sl, dl; |
| 51 | |
| 52 | if((i = iconv_open(to, from)) == (iconv_t)-1) |
| 53 | fatal(errno, "error calling iconv_open"); |
| 54 | do { |
| 55 | bufsize = bufsize ? 2 * bufsize : 32; |
| 56 | buf = xrealloc_noptr(buf, bufsize); |
| 57 | iconv(i, 0, 0, 0, 0); |
| 58 | s = (char *)ptr; |
| 59 | sl = n; |
| 60 | d = buf; |
| 61 | dl = bufsize; |
| 62 | /* (void *) to work around FreeBSD's nonstandard iconv prototype */ |
| 63 | len = iconv(i, (void *)&s, &sl, &d, &dl); |
| 64 | } while(len == (size_t)-1 && errno == E2BIG); |
| 65 | iconv_close(i); |
| 66 | if(len == (size_t)-1) { |
| 67 | error(errno, "error converting from %s to %s", from, to); |
| 68 | return 0; |
| 69 | } |
| 70 | return buf; |
| 71 | } |
| 72 | |
| 73 | /** @brief Convert UTF-8 to UCS-4 |
| 74 | * @param mb Pointer to 0-terminated UTF-8 string |
| 75 | * @return Pointer to 0-terminated UCS-4 string |
| 76 | * |
| 77 | * Not everybody's iconv supports UCS-4, and it's inconvenient to have to know |
| 78 | * our endianness, and it's easy to convert it ourselves, so we do. See also |
| 79 | * @ref ucs42utf8(). |
| 80 | */ |
| 81 | uint32_t *utf82ucs4(const char *mb) { |
| 82 | struct dynstr_ucs4 d; |
| 83 | uint32_t c; |
| 84 | |
| 85 | dynstr_ucs4_init(&d); |
| 86 | while(*mb) { |
| 87 | PARSE_UTF8(mb, c, |
| 88 | error(0, "invalid UTF-8 sequence"); return 0;); |
| 89 | dynstr_ucs4_append(&d, c); |
| 90 | } |
| 91 | dynstr_ucs4_terminate(&d); |
| 92 | return d.vec; |
| 93 | } |
| 94 | |
| 95 | /** @brief Convert one UCS-4 character to UTF-8 |
| 96 | * @param c Character to convert |
| 97 | * @param d Dynamic string to append UTF-8 sequence to |
| 98 | * @return 0 on success, -1 on error |
| 99 | */ |
| 100 | int one_ucs42utf8(uint32_t c, struct dynstr *d) { |
| 101 | if(c < 0x80) |
| 102 | dynstr_append(d, c); |
| 103 | else if(c < 0x800) { |
| 104 | dynstr_append(d, 0xC0 | (c >> 6)); |
| 105 | dynstr_append(d, 0x80 | (c & 0x3F)); |
| 106 | } else if(c < 0x10000) { |
| 107 | dynstr_append(d, 0xE0 | (c >> 12)); |
| 108 | dynstr_append(d, 0x80 | ((c >> 6) & 0x3F)); |
| 109 | dynstr_append(d, 0x80 | (c & 0x3F)); |
| 110 | } else if(c < 0x110000) { |
| 111 | dynstr_append(d, 0xF0 | (c >> 18)); |
| 112 | dynstr_append(d, 0x80 | ((c >> 12) & 0x3F)); |
| 113 | dynstr_append(d, 0x80 | ((c >> 6) & 0x3F)); |
| 114 | dynstr_append(d, 0x80 | (c & 0x3F)); |
| 115 | } else { |
| 116 | error(0, "invalid UCS-4 character %#"PRIx32, c); |
| 117 | return -1; |
| 118 | } |
| 119 | return 0; |
| 120 | } |
| 121 | |
| 122 | /** @brief Convert UCS-4 to UTF-8 |
| 123 | * @param u Pointer to 0-terminated UCS-4 string |
| 124 | * @return Pointer to 0-terminated UTF-8 string |
| 125 | * |
| 126 | * See @ref utf82ucs4(). |
| 127 | */ |
| 128 | char *ucs42utf8(const uint32_t *u) { |
| 129 | struct dynstr d; |
| 130 | uint32_t c; |
| 131 | |
| 132 | dynstr_init(&d); |
| 133 | while((c = *u++)) { |
| 134 | if(one_ucs42utf8(c, &d)) |
| 135 | return 0; |
| 136 | } |
| 137 | dynstr_terminate(&d); |
| 138 | return d.vec; |
| 139 | } |
| 140 | |
| 141 | /** @brief Convert from the local multibyte encoding to UTF-8 */ |
| 142 | char *mb2utf8(const char *mb) { |
| 143 | return convert(nl_langinfo(CODESET), "UTF-8", mb, strlen(mb) + 1); |
| 144 | } |
| 145 | |
| 146 | /** @brief Convert from UTF-8 to the local multibyte encoding */ |
| 147 | char *utf82mb(const char *utf8) { |
| 148 | return convert("UTF-8", nl_langinfo(CODESET), utf8, strlen(utf8) + 1); |
| 149 | } |
| 150 | |
| 151 | /** @brief Convert from encoding @p from to UTF-8 */ |
| 152 | char *any2utf8(const char *from, const char *any) { |
| 153 | return convert(from, "UTF-8", any, strlen(any) + 1); |
| 154 | } |
| 155 | |
| 156 | /** @brief Convert from encoding @p from to the local multibyte encoding */ |
| 157 | char *any2mb(const char *from, const char *any) { |
| 158 | if(from) return convert(from, nl_langinfo(CODESET), any, strlen(any) + 1); |
| 159 | else return xstrdup(any); |
| 160 | } |
| 161 | |
| 162 | /** @brief Convert from encoding @p from to encoding @p to */ |
| 163 | char *any2any(const char *from, |
| 164 | const char *to, |
| 165 | const char *any) { |
| 166 | if(from || to) return convert(from, to, any, strlen(any) + 1); |
| 167 | else return xstrdup(any); |
| 168 | } |
| 169 | |
| 170 | /** @brief strlen workalike for UCS-4 strings |
| 171 | * |
| 172 | * We don't rely on the local @c wchar_t being UCS-4. |
| 173 | */ |
| 174 | int ucs4cmp(const uint32_t *a, const uint32_t *b) { |
| 175 | while(*a && *b && *a == *b) ++a, ++b; |
| 176 | if(*a > *b) return 1; |
| 177 | else if(*a < *b) return -1; |
| 178 | else return 0; |
| 179 | } |
| 180 | |
| 181 | /** @brief Return nonzero if @p c is a combining character */ |
| 182 | static int combining(int c) { |
| 183 | if(c < UNICODE_NCHARS) { |
| 184 | const struct unidata *const ud = &unidata[c / 256][c % 256]; |
| 185 | |
| 186 | return ud->gc == unicode_gc_Mn || ud->ccc != 0; |
| 187 | } |
| 188 | /* Assume unknown characters are noncombining */ |
| 189 | return 0; |
| 190 | } |
| 191 | |
| 192 | /** @brief Truncate a string for display purposes |
| 193 | * @param s Pointer to UTF-8 string |
| 194 | * @param max Maximum number of columns |
| 195 | * @return @p or truncated string (never NULL) |
| 196 | * |
| 197 | * We don't correctly support bidi or double-width characters yet, nor |
| 198 | * locate default grapheme cluster boundaries for saner truncation. |
| 199 | */ |
| 200 | const char *truncate_for_display(const char *s, long max) { |
| 201 | const char *t = s, *r, *cut = 0; |
| 202 | char *truncated; |
| 203 | uint32_t c; |
| 204 | long n = 0; |
| 205 | |
| 206 | /* We need to discover two things: firstly whether the string is |
| 207 | * longer than @p max glyphs and secondly if it is not, where to cut |
| 208 | * the string. |
| 209 | * |
| 210 | * Combining characters follow their base character (unicode |
| 211 | * standard 5.0 s2.11), so after each base character we must |
| 212 | */ |
| 213 | while(*t) { |
| 214 | PARSE_UTF8(t, c, return s); |
| 215 | if(combining(c)) |
| 216 | /* This must be an initial combining character. We just skip it. */ |
| 217 | continue; |
| 218 | /* So c must be a base character. It may be followed by any |
| 219 | * number of combining characters. We advance past them. */ |
| 220 | do { |
| 221 | r = t; |
| 222 | PARSE_UTF8(t, c, return s); |
| 223 | } while(combining(c)); |
| 224 | /* Last character wasn't a combining character so back up */ |
| 225 | t = r; |
| 226 | ++n; |
| 227 | /* So now there are N glyphs before position T. We might |
| 228 | * therefore have reached the cut position. */ |
| 229 | if(n == max - 3) |
| 230 | cut = t; |
| 231 | } |
| 232 | /* If the string is short enough we return it unmodified */ |
| 233 | if(n < max) |
| 234 | return s; |
| 235 | truncated = xmalloc_noptr(cut - s + 4); |
| 236 | memcpy(truncated, s, cut - s); |
| 237 | strcpy(truncated + (cut - s), "..."); |
| 238 | return truncated; |
| 239 | } |
| 240 | |
| 241 | /* |
| 242 | Local Variables: |
| 243 | c-basic-offset:2 |
| 244 | comment-column:40 |
| 245 | End: |
| 246 | */ |