src/intl/charsets.h

   1 #ifndef EL__INTL_CHARSETS_H
   2 #define EL__INTL_CHARSETS_H
   3
   4 typedef uint32_t unicode_val_T;
   5
   6 /* U+0020 SPACE.  Normally the same as ' ' or L' ' but perhaps ELinks
   7  * shouldn't rely on that.  */
   8 #define UCS_SPACE ((unicode_val_T) 0x0020)
   9
  10 /* U+00A0 NO-BREAK SPACE.  */
  11 #define UCS_NO_BREAK_SPACE ((unicode_val_T) 0x00A0)
  12
  13 /* U+00AD SOFT HYPHEN.  */
  14 #define UCS_SOFT_HYPHEN ((unicode_val_T) 0x00AD)
  15
  16 /* U+FFFD REPLACEMENT CHARACTER.  Used when no Unicode mapping is
  17  * known for a byte in a codepage, or when invalid UTF-8 is received
  18  * from a terminal.  After generating the character, ELinks then
  19  * treats it like any other Unicode character.  The user can also type
  20  * this character directly, and it can occur in documents.  */
  21 #define UCS_REPLACEMENT_CHARACTER ((unicode_val_T) 0xFFFD)
  22
  23 /* A special value that fits in unicode_val_T but is outside the range
  24  * of Unicode characters.  utf8_to_unicode and cp_to_unicode return
  25  * this if the input is too short.  This is also used as a placeholder
  26  * for the second cell of a double-cell character.  */
  27 #define UCS_NO_CHAR ((unicode_val_T) 0xFFFFFFFD)
  28
  29 /* If ELinks should display a double-cell character but there is only
  30  * one cell available, it displays this character instead.  This must
  31  * be a single-cell character but need not be unique.  Possible values
  32  * might be U+0020 SPACE or U+303F IDEOGRAPHIC HALF FILL SPACE.
  33  *
  34  * Some BFU widgets (at least input fields and list boxes) currently
  35  * ignore this setting and use U+0020 instead.  (They first draw spaces
  36  * and then overwrite with text; look for utf8_cells2bytes calls.)
  37  * We should fix that if we ever change the value.  */
  38 #define UCS_ORPHAN_CELL ((unicode_val_T) 0x20)
  39
  40 /* &nbsp; replacement character. See u2cp().
  41  * UTF-8 strings should use the encoding of U+00A0 instead. */
  42 #define NBSP_CHAR ((unsigned char) 1)
  43 #define NBSP_CHAR_STRING "\001"
  44
  45 /* How to convert a byte from a source charset.  This is used in an
  46  * array (struct conv_table[256]) indexed by the byte value.  */
  47 struct conv_table {
  48         /* 0 if this is the final byte of a character, or 1 if more
  49          * bytes are needed.  */
  50         int t;
  51         union {
  52                 /* If @t==0: a null-terminated string that is the
  53                  * corresponding character in the target charset.
  54                  * Normally, the string is statically allocated.
  55                  * However, if the conversion table is to UTF-8, then
  56                  * the strings in elements 0x80 to 0xFF are allocated
  57                  * with @mem_alloc and owned by the table.  */
  58                 const unsigned char *str;
  59                 /* If @t==1: a pointer to a nested conversion table
  60                  * (with 256 elements) that describes how to convert
  61                  * each possible subsequent byte.  The conversion
  62                  * table owns the nested conversion table.  */
  63                 struct conv_table *tbl;
  64         } u;
  65 };
  66
  67 enum convert_string_mode {
  68         CSM_DEFAULT, /* Convert any char. */
  69         CSM_QUERY, /* Special handling of '&' and '=' chars. */
  70         CSM_FORM, /* Special handling of '&' and '=' chars in forms. */
  71         CSM_NONE, /* Convert nothing. */
  72 };
  73
  74 /* How to translate U+00A0 NO-BREAK SPACE.  If u2cp_ is converting to
  75  * UTF-8, it ignores this choice and just encodes the U+00A0.  */
  76 enum nbsp_mode {
  77         /* Convert to NBSP_CHAR.  This lets the HTML renderer
  78          * recognize nbsp even if the codepage doesn't support
  79          * nbsp.  (VISCII doesn't.)  */
  80         NBSP_MODE_HACK = 0,
  81
  82         /* Convert to normal ASCII space.  */
  83         NBSP_MODE_ASCII = 1
  84 };
  85
  86 struct conv_table *get_translation_table(int, int);
  87 const unsigned char *get_entity_string(const unsigned char *str,
  88                                        const int strlen, int encoding);
  89
  90 /* The convert_string() name is also used by Samba (version 3.0.3), which
  91  * provides libnss_wins.so.2, which is called somewhere inside
  92  * _nss_wins_gethostbyname_r(). This name clash causes the elinks hostname
  93  * lookup thread to crash so we need to rename the symbol. */
  94 /* Reported by Derek Poon and filed as bug 453 */
  95 #undef convert_string
  96 #define convert_string convert_string_elinks
  97
  98 /* This routine converts a string from one charset to another according to the
  99  * passed @convert_table, potentially also decoding SGML (HTML) entities along
 100  * the way (according to @mode). It either returns dynamically allocated
 101  * converted string of length @length, or if the @callback is non-NULL it calls
 102  * it each few bytes instead and always returns NULL (@length is undefined).
 103  * Note that it's ok not to care and pass NULL as @length. */
 104 unsigned char *convert_string(struct conv_table *convert_table,
 105                               unsigned char *chars, int charslen, int cp,
 106                               enum convert_string_mode mode, int *length,
 107                               void (*callback)(void *data, unsigned char *buf, int buflen),
 108                               void *callback_data);
 109
 110 int get_cp_index(unsigned char *);
 111 unsigned char *get_cp_name(int);
 112 unsigned char *get_cp_config_name(int);
 113 unsigned char *get_cp_mime_name(int);
 114 int is_cp_utf8(int);
 115 void free_conv_table(void);
 116 #ifdef CONFIG_UTF8
 117 inline unsigned char *encode_utf8(unicode_val_T);
 118 inline unsigned char *utf8_prevchar(unsigned char *, int, unsigned char *);
 119 inline int utf8charlen(const unsigned char *);
 120 int utf8_char2cells(unsigned char *, unsigned char *);
 121 int utf8_ptr2cells(unsigned char *, unsigned char *);
 122 int utf8_ptr2chars(unsigned char *, unsigned char *);
 123 int utf8_cells2bytes(unsigned char *, int, unsigned char *);
 124 /* How utf8_step_forward and utf8_step_backward count steps.  */
 125 enum utf8_step {
 126         /* Each step is one character, even if it is a combining or
 127          * double-width character.  */
 128         UTF8_STEP_CHARACTERS,
 129
 130         /* Each step is one cell.  If the specified number of steps
 131          * would end in the middle of a double-width character, do not
 132          * include the character.  */
 133         UTF8_STEP_CELLS_FEWER,
 134
 135         /* Each step is one cell.  If the specified number of steps
 136          * would end in the middle of a double-width character,
 137          * include the whole character.  */
 138         UTF8_STEP_CELLS_MORE
 139 };
 140 unsigned char *utf8_step_forward(unsigned char *, unsigned char *,
 141                                  int, enum utf8_step, int *);
 142 unsigned char *utf8_step_backward(unsigned char *, unsigned char *,
 143                                   int, enum utf8_step, int *);
 144 inline int unicode_to_cell(unicode_val_T);
 145 unicode_val_T unicode_fold_label_case(unicode_val_T);
 146 inline int strlen_utf8(unsigned char **);
 147 inline unicode_val_T utf8_to_unicode(unsigned char **, const unsigned char *);
 148 unicode_val_T cp_to_unicode(int, unsigned char **, unsigned char *);
 149 #endif /* CONFIG_UTF8 */
 150
 151 unicode_val_T cp2u(int, unsigned char);
 152 const unsigned char *cp2utf8(int, int);
 153
 154 const unsigned char *u2cp_(unicode_val_T, int, enum nbsp_mode);
 155 #define u2cp(u, to) u2cp_(u, to, NBSP_MODE_HACK)
 156 #define u2cp_no_nbsp(u, to) u2cp_(u, to, NBSP_MODE_ASCII)
 157
 158 void init_charsets_lookup(void);
 159 void free_charsets_lookup(void);
 160
 161 /* UTF-16 encodes each Unicode character U+0000...U+FFFF as a single
 162  * 16-bit code unit, and each character U+10000...U+10FFFF as a pair
 163  * of two code units: a high surrogate followed by a low surrogate.
 164  * The range U+D800...U+DFFF is reserved for these surrogates.  */
 165 #define is_utf16_surrogate(u)           (((u) & 0xFFFFF800) == 0xD800)
 166 #define is_utf16_high_surrogate(u)      (((u) & 0xFFFFFC00) == 0xD800)
 167 #define is_utf16_low_surrogate(u)       (((u) & 0xFFFFFC00) == 0xDC00)
 168 #define join_utf16_surrogates(high,low) (0x10000 + (((high) - 0xD800L) << 10) + ((low) - 0xDC00))
 169 #define needs_utf16_surrogates(u)       ((uint32_t) ((u) - 0x10000) < 0x100000)
 170 #define get_utf16_high_surrogate(u)     (0xD800 + (((u) - 0x10000) >> 10))
 171 #define get_utf16_low_surrogate(u)      (0xDC00 + ((u) & 0x3FF))
 172
 173 #endif