/* Basic multilingual character support.
- Copyright (C) 1995, 1997 Electrotechnical Laboratory, JAPAN.
+ Copyright (C) 1995, 1997, 1998 Electrotechnical Laboratory, JAPAN.
Licensed to the Free Software Foundation.
This file is part of GNU Emacs.
/* At first, see the document in `charset.h' to understand the code in
this file. */
+#ifdef emacs
+#include <config.h>
+#endif
+
#include <stdio.h>
#ifdef emacs
#include <sys/types.h>
-#include <config.h>
#include "lisp.h"
#include "buffer.h"
#include "charset.h"
#endif /* emacs */
-Lisp_Object Qcharset, Qascii, Qcomposition;
+Lisp_Object Qcharset, Qascii;
+Lisp_Object Qunknown;
/* Declaration of special leading-codes. */
-int leading_code_composition; /* for composite characters */
int leading_code_private_11; /* for private DIMENSION1 of 1-column */
int leading_code_private_12; /* for private DIMENSION1 of 2-column */
int leading_code_private_21; /* for private DIMENSION2 of 1-column */
/* Declaration of special charsets. */
int charset_ascii; /* ASCII */
-int charset_composition; /* for a composite character */
int charset_latin_iso8859_1; /* ISO8859-1 (Latin-1) */
int charset_jisx0208_1978; /* JISX0208.1978 (Japanese Kanji old set) */
int charset_jisx0208; /* JISX0208.1983 (Japanese Kanji) */
/* A list of charset symbols ever defined. */
Lisp_Object Vcharset_list;
+/* Vector of translation table ever defined.
+ ID of a translation table is used to index this vector. */
+Lisp_Object Vtranslation_table_vector;
+
+/* A char-table for characters which may invoke auto-filling. */
+Lisp_Object Vauto_fill_chars;
+
+Lisp_Object Qauto_fill_chars;
+
/* Tables used by macros BYTES_BY_CHAR_HEAD and WIDTH_BY_CHAR_HEAD. */
int bytes_by_char_head[256];
int width_by_char_head[256];
CHARS, and FINAL-CHAR) to Emacs' charset. */
int iso_charset_table[2][2][128];
-/* Table of pointers to the structure `cmpchar_info' indexed by
- CMPCHAR-ID. */
-struct cmpchar_info **cmpchar_table;
-/* The current size of `cmpchar_table'. */
-static int cmpchar_table_size;
-/* Number of the current composite characters. */
-int n_cmpchars;
-
/* Variables used locally in the macro FETCH_MULTIBYTE_CHAR. */
unsigned char *_fetch_multibyte_char_p;
int _fetch_multibyte_char_len;
-/* Set STR a pointer to the multi-byte form of the character C. If C
- is not a composite character, the multi-byte form is set in WORKBUF
- and STR points WORKBUF. The caller should allocate at least 4-byte
- area at WORKBUF in advance. Returns the length of the multi-byte
- form. If C is an invalid character to have a multi-byte form,
+/* Offset to add to a non-ASCII value when inserting it. */
+int nonascii_insert_offset;
+
+/* Translation table for converting non-ASCII unibyte characters
+ to multibyte codes, or nil. */
+Lisp_Object Vnonascii_translation_table;
+
+/* List of all possible generic characters. */
+Lisp_Object Vgeneric_character_list;
+
+#define min(X, Y) ((X) < (Y) ? (X) : (Y))
+#define max(X, Y) ((X) > (Y) ? (X) : (Y))
+\f
+void
+invalid_character (c)
+ int c;
+{
+ error ("Invalid character: 0%o, %d, 0x%x", c, c, c);
+}
+
+/* Parse a multibyte character string STR of length LENGTH (>= 2) set
+ BYTES to the length of actual multibyte sequence, CHARSET, C1, and
+ C2 to such values that MAKE_CHAR can make the multibyte character
+ from them.
+
+ It is assumed that *STR is one of base leading codes and the
+ following (LENGTH - 1) bytes satisfy !CHAR_HEAD_P.
+
+ This macro should be called only from SPLIT_MULTIBYTE_SEQ. */
+
+#define SPLIT_CHARACTER_SEQ(str, length, bytes, charset, c1, c2) \
+ do { \
+ (bytes) = 1; \
+ (charset) = (str)[0]; \
+ if ((charset) >= LEADING_CODE_PRIVATE_11 \
+ && (charset) <= LEADING_CODE_PRIVATE_22) \
+ (charset) = (str)[(bytes)++]; \
+ if ((bytes) < (length)) \
+ { \
+ (c1) = (str)[(bytes)++] & 0x7F; \
+ if ((bytes) < (length)) \
+ (c2) = (str)[(bytes)++] & 0x7F; \
+ else \
+ (c2) = -1; \
+ } \
+ else \
+ (c1) = (c2) = -1; \
+ } while (0)
+
+/* Parse string STR of length LENGTH and check if a multibyte
+ characters is at STR. Set BYTES to the actual length, CHARSET, C1,
+ C2 to proper values for that character. */
+
+#define SPLIT_MULTIBYTE_SEQ(str, length, bytes, charset, c1, c2) \
+ do { \
+ int i; \
+ if (ASCII_BYTE_P ((str)[0])) \
+ i = 1; \
+ else \
+ for (i = 1; i < (length) && ! CHAR_HEAD_P ((str)[i]); i++); \
+ if (i == 1) \
+ (bytes) = 1, (charset) = CHARSET_ASCII, (c1) = (str)[0] ; \
+ else \
+ { \
+ if (i > BYTES_BY_CHAR_HEAD ((str)[0])) \
+ i = BYTES_BY_CHAR_HEAD ((str)[0]); \
+ SPLIT_CHARACTER_SEQ (str, i, bytes, charset, c1, c2); \
+ } \
+ } while (0)
+
+/* 1 if CHARSET, C1, and C2 compose a valid character, else 0. */
+#define CHAR_COMPONENTS_VALID_P(charset, c1, c2) \
+ (charset == CHARSET_ASCII \
+ ? ((c1) >= 0 && (c1) <= 0x7F) \
+ : (CHARSET_DIMENSION (charset) == 1 \
+ ? ((c1) >= 0x20 && (c1) <= 0x7F) \
+ : ((c1) >= 0x20 && (c1) <= 0x7F && (c2) >= 0x20 && (c2) <= 0x7F)))
+
+/* Store multi-byte form of the character C in STR. The caller should
+ allocate at least 4-byte area at STR in advance. Returns the
+ length of the multi-byte form. If C is an invalid character code,
signal an error.
- Use macro `CHAR_STRING (C, WORKBUF, STR)' instead of calling this
- function directly if C can be an ASCII character. */
+ Use macro `CHAR_STRING (C, STR)' instead of calling this function
+ directly if C can be an ASCII character. */
int
-non_ascii_char_to_string (c, workbuf, str)
+char_to_string (c, str)
int c;
- unsigned char *workbuf, **str;
+ unsigned char *str;
{
- int charset, c1, c2;
+ unsigned char *p = str;
- if (COMPOSITE_CHAR_P (c))
+ if (c & CHAR_MODIFIER_MASK) /* This includes the case C is negative. */
{
- int cmpchar_id = COMPOSITE_CHAR_ID (c);
+ /* Multibyte character can't have a modifier bit. */
+ if (! SINGLE_BYTE_CHAR_P ((c & ~CHAR_MODIFIER_MASK)))
+ invalid_character (c);
- if (cmpchar_id < n_cmpchars)
+ /* For Meta, Shift, and Control modifiers, we need special care. */
+ if (c & CHAR_META)
{
- *str = cmpchar_table[cmpchar_id]->data;
- return cmpchar_table[cmpchar_id]->len;
+ /* Move the meta bit to the right place for a string. */
+ c = (c & ~CHAR_META) | 0x80;
}
- else
+ if (c & CHAR_SHIFT)
+ {
+ /* Shift modifier is valid only with [A-Za-z]. */
+ if ((c & 0377) >= 'A' && (c & 0377) <= 'Z')
+ c &= ~CHAR_SHIFT;
+ else if ((c & 0377) >= 'a' && (c & 0377) <= 'z')
+ c = (c & ~CHAR_SHIFT) - ('a' - 'A');
+ }
+ if (c & CHAR_CTL)
+ {
+ /* Simulate the code in lread.c. */
+ /* Allow `\C- ' and `\C-?'. */
+ if (c == (CHAR_CTL | ' '))
+ c = 0;
+ else if (c == (CHAR_CTL | '?'))
+ c = 127;
+ /* ASCII control chars are made from letters (both cases),
+ as well as the non-letters within 0100...0137. */
+ else if ((c & 0137) >= 0101 && (c & 0137) <= 0132)
+ c &= (037 | (~0177 & ~CHAR_CTL));
+ else if ((c & 0177) >= 0100 && (c & 0177) <= 0137)
+ c &= (037 | (~0177 & ~CHAR_CTL));
+ }
+
+ /* If C still has any modifier bits, it is an invalid character. */
+ if (c & CHAR_MODIFIER_MASK)
+ invalid_character (c);
+
+ *p++ = c;
+ }
+ else if (c < MAX_CHAR)
+ {
+ int charset, c1, c2;
+
+ SPLIT_NON_ASCII_CHAR (c, charset, c1, c2);
+
+ if (charset >= LEADING_CODE_EXT_11)
+ *p++ = (charset < LEADING_CODE_EXT_12
+ ? LEADING_CODE_PRIVATE_11
+ : (charset < LEADING_CODE_EXT_21
+ ? LEADING_CODE_PRIVATE_12
+ : (charset < LEADING_CODE_EXT_22
+ ? LEADING_CODE_PRIVATE_21
+ : LEADING_CODE_PRIVATE_22)));
+ *p++ = charset;
+ if (c1 > 0 && c1 < 32 || c2 > 0 && c2 < 32)
+ invalid_character (c);
+ if (c1)
{
- error ("Invalid character: %d", c);
+ *p++ = c1 | 0x80;
+ if (c2 > 0)
+ *p++ = c2 | 0x80;
}
}
- SPLIT_NON_ASCII_CHAR (c, charset, c1, c2);
- if (!charset
- || ! CHARSET_DEFINED_P (charset)
- || c1 >= 0 && c1 < 32
- || c2 >= 0 && c2 < 32)
- error ("Invalid character: %d", c);
-
- *str = workbuf;
- *workbuf++ = CHARSET_LEADING_CODE_BASE (charset);
- if (*workbuf = CHARSET_LEADING_CODE_EXT (charset))
- workbuf++;
- *workbuf++ = c1 | 0x80;
- if (c2 >= 0)
- *workbuf++ = c2 | 0x80;
-
- return (workbuf - *str);
+ return (p -str);
}
-/* Return a non-ASCII character of which multi-byte form is at STR of
- length LEN. If ACTUAL_LEN is not NULL, the actual length of the
- character is set to the address ACTUAL_LEN.
+/* Return the non-ASCII character corresponding to multi-byte form at
+ STR of length LEN. If ACTUAL_LEN is not NULL, store the byte
+ length of the multibyte form in *ACTUAL_LEN.
- Use macro `STRING_CHAR (STR, LEN)' instead of calling this function
- directly if STR can hold an ASCII character. */
+ Use macros STRING_CHAR or STRING_CHAR_AND_LENGTH instead of calling
+ this function directly if you want ot handle ASCII characters as
+ well. */
-string_to_non_ascii_char (str, len, actual_len)
- unsigned char *str;
+int
+string_to_char (str, len, actual_len)
+ const unsigned char *str;
int len, *actual_len;
{
- int charset;
- unsigned char c1, c2;
- register int c;
-
- if (SPLIT_STRING (str, len, charset, c1, c2) == CHARSET_ASCII)
- {
- if (actual_len)
- *actual_len = 1;
- return (int) *str;
- }
-
- c = MAKE_NON_ASCII_CHAR (charset, c1, c2);
+ int c, bytes, charset, c1, c2;
+ SPLIT_MULTIBYTE_SEQ (str, len, bytes, charset, c1, c2);
+ c = MAKE_CHAR (charset, c1, c2);
if (actual_len)
- *actual_len = (charset == CHARSET_COMPOSITION
- ? cmpchar_table[COMPOSITE_CHAR_ID (c)]->len
- : BYTES_BY_CHAR_HEAD (*str));
+ *actual_len = bytes;
return c;
}
-/* Return the length of the multi-byte form at string STR of length LEN. */
+/* Return the length of the multi-byte form at string STR of length LEN.
+ Use the macro MULTIBYTE_FORM_LENGTH instead. */
int
multibyte_form_length (str, len)
- unsigned char *str;
+ const unsigned char *str;
int len;
{
- int charset;
- unsigned char c1, c2;
- register int c;
-
- if (SPLIT_STRING (str, len, charset, c1, c2) == CHARSET_ASCII)
- return 1;
+ int bytes;
- return (charset == CHARSET_COMPOSITION
- ? cmpchar_table[(c1 << 7) | c2]->len
- : BYTES_BY_CHAR_HEAD (*str));
+ PARSE_MULTIBYTE_SEQ (str, len, bytes);
+ return bytes;
}
-/* Check if string STR of length LEN contains valid multi-byte form of
- a character. If valid, charset and position codes of the character
- is set at *CHARSET, *C1, and *C2, and return 0. If not valid,
+/* Check multibyte form at string STR of length LEN and set variables
+ pointed by CHARSET, C1, and C2 to charset and position codes of the
+ character at STR, and return 0. If there's no multibyte character,
return -1. This should be used only in the macro SPLIT_STRING
which checks range of STR in advance. */
-split_non_ascii_string (str, len, charset, c1, c2)
- register unsigned char *str, *c1, *c2;
- register int len, *charset;
+int
+split_string (str, len, charset, c1, c2)
+ const unsigned char *str;
+ unsigned char *c1, *c2;
+ int len, *charset;
{
- register unsigned int cs = *str++;
-
- if (cs == LEADING_CODE_COMPOSITION)
- {
- int cmpchar_id = str_cmpchar_id (str - 1, len);
+ register int bytes, cs, code1, code2 = -1;
- if (cmpchar_id < 0)
- return -1;
- *charset = cs, *c1 = cmpchar_id >> 7, *c2 = cmpchar_id & 0x7F;
- }
- else if ((cs < LEADING_CODE_PRIVATE_11 || (cs = *str++) >= 0xA0)
- && CHARSET_DEFINED_P (cs))
- {
- *charset = cs;
- if (*str < 0xA0)
- return -1;
- *c1 = (*str++) & 0x7F;
- if (CHARSET_DIMENSION (cs) == 2)
- {
- if (*str < 0xA0)
- return -1;
- *c2 = (*str++) & 0x7F;
- }
- }
- else
+ SPLIT_MULTIBYTE_SEQ (str, len, bytes, cs, code1, code2);
+ if (cs == CHARSET_ASCII)
return -1;
+ *charset = cs;
+ *c1 = code1;
+ *c2 = code2;
return 0;
}
-/* Return a character unified with C (or a character made of CHARSET,
- C1, and C2) in unification table TABLE. If no unification is found
- in TABLE, return C. */
-unify_char (table, c, charset, c1, c2)
+/* Return 1 iff character C has valid printable glyph.
+ Use the macro CHAR_PRINTABLE_P instead. */
+int
+char_printable_p (c)
+ int c;
+{
+ int charset, c1, c2, chars;
+
+ if (SINGLE_BYTE_CHAR_P (c))
+ return 1;
+ if (c >= MAX_CHAR)
+ return 0;
+
+ SPLIT_NON_ASCII_CHAR (c, charset, c1, c2);
+ if (! CHARSET_DEFINED_P (charset))
+ return 0;
+ if (CHARSET_CHARS (charset) == 94
+ ? c1 <= 32 || c1 >= 127
+ : c1 < 32)
+ return 0;
+ if (CHARSET_DIMENSION (charset) == 2
+ && (CHARSET_CHARS (charset) == 94
+ ? c2 <= 32 || c2 >= 127
+ : c2 < 32))
+ return 0;
+ return 1;
+}
+
+/* Translate character C by translation table TABLE. If C
+ is negative, translate a character specified by CHARSET, C1, and C2
+ (C1 and C2 are code points of the character). If no translation is
+ found in TABLE, return C. */
+int
+translate_char (table, c, charset, c1, c2)
Lisp_Object table;
int c, charset, c1, c2;
{
Lisp_Object ch;
int alt_charset, alt_c1, alt_c2, dimension;
- if (c < 0) c = MAKE_CHAR (charset, c1, c2);
+ if (c < 0) c = MAKE_CHAR (charset, (c1 & 0x7F) , (c2 & 0x7F));
if (!CHAR_TABLE_P (table)
- || (ch = Faref (table, make_number (c)), !INTEGERP (ch))
- || XINT (ch) < 0)
+ || (ch = Faref (table, make_number (c)), !NATNUMP (ch)))
return c;
SPLIT_CHAR (XFASTINT (ch), alt_charset, alt_c1, alt_c2);
if (dimension != CHARSET_DIMENSION (charset))
/* We can't make such a character because of dimension mismatch. */
return c;
- if (!alt_c1) alt_c1 = c1;
- if (!alt_c2) alt_c2 = c2;
return MAKE_CHAR (alt_charset, c1, c2);
}
+/* Convert the unibyte character C to multibyte based on
+ Vnonascii_translation_table or nonascii_insert_offset. If they can't
+ convert C to a valid multibyte character, convert it based on
+ DEFAULT_NONASCII_INSERT_OFFSET which makes C a Latin-1 character. */
+
+int
+unibyte_char_to_multibyte (c)
+ int c;
+{
+ if (c < 0400 && c >= 0200)
+ {
+ int c_save = c;
+
+ if (! NILP (Vnonascii_translation_table))
+ {
+ c = XINT (Faref (Vnonascii_translation_table, make_number (c)));
+ if (c >= 0400 && ! char_valid_p (c, 0))
+ c = c_save + DEFAULT_NONASCII_INSERT_OFFSET;
+ }
+ else if (c >= 0240 && nonascii_insert_offset > 0)
+ {
+ c += nonascii_insert_offset;
+ if (c < 0400 || ! char_valid_p (c, 0))
+ c = c_save + DEFAULT_NONASCII_INSERT_OFFSET;
+ }
+ else if (c >= 0240)
+ c = c_save + DEFAULT_NONASCII_INSERT_OFFSET;
+ }
+ return c;
+}
+
+
+/* Convert the multibyte character C to unibyte 8-bit character based
+ on Vnonascii_translation_table or nonascii_insert_offset. If
+ REV_TBL is non-nil, it should be a reverse table of
+ Vnonascii_translation_table, i.e. what given by:
+ Fchar_table_extra_slot (Vnonascii_translation_table, make_number (0)) */
+
+int
+multibyte_char_to_unibyte (c, rev_tbl)
+ int c;
+ Lisp_Object rev_tbl;
+{
+ if (!SINGLE_BYTE_CHAR_P (c))
+ {
+ int c_save = c;
+
+ if (! CHAR_TABLE_P (rev_tbl)
+ && CHAR_TABLE_P (Vnonascii_translation_table))
+ rev_tbl = Fchar_table_extra_slot (Vnonascii_translation_table,
+ make_number (0));
+ if (CHAR_TABLE_P (rev_tbl))
+ {
+ Lisp_Object temp;
+ temp = Faref (rev_tbl, make_number (c));
+ if (INTEGERP (temp))
+ c = XINT (temp);
+ if (c >= 256)
+ c = (c_save & 0177) + 0200;
+ }
+ else
+ {
+ if (nonascii_insert_offset > 0)
+ c -= nonascii_insert_offset;
+ if (c < 128 || c >= 256)
+ c = (c_save & 0177) + 0200;
+ }
+ }
+
+ return c;
+}
+
+\f
/* Update the table Vcharset_table with the given arguments (see the
document of `define-charset' for the meaning of each argument).
Several other table contents are also updated. The caller should
leading_code_ext = charset;
}
+ if (BYTES_BY_CHAR_HEAD (leading_code_base) != bytes)
+ error ("Invalid dimension for the charset-ID %d", charset);
+
CHARSET_TABLE_INFO (charset, CHARSET_ID_IDX) = charset_id;
CHARSET_TABLE_INFO (charset, CHARSET_BYTES_IDX) = make_number (bytes);
CHARSET_TABLE_INFO (charset, CHARSET_DIMENSION_IDX) = dimension;
if (charset != CHARSET_ASCII
&& charset < MIN_CHARSET_PRIVATE_DIMENSION1)
{
- /* Update tables bytes_by_char_head and width_by_char_head. */
- bytes_by_char_head[leading_code_base] = bytes;
width_by_char_head[leading_code_base] = XINT (width);
/* Update table emacs_code_class. */
if (width == 1)
from = LEADING_CODE_EXT_21, to = LEADING_CODE_EXT_22;
else
- from = LEADING_CODE_EXT_22, to = LEADING_CODE_EXT_MAX - 1;
+ from = LEADING_CODE_EXT_22, to = LEADING_CODE_EXT_MAX + 1;
}
for (charset = from; charset < to; charset++)
occupies on the screen: one of 0, 1, and 2.\n\
\n\
DIRECTION (integer) is the rendering direction of characters in the\n\
-charset when rendering. If 0, render from right to left, else\n\
-render from left to right.\n\
+charset when rendering. If 0, render from left to right, else\n\
+render from right to left.\n\
\n\
ISO-FINAL-CHAR (character) is the final character of the\n\
corresponding ISO 2022 charset.\n\
return Qnil;
}
+DEFUN ("generic-character-list", Fgeneric_character_list,
+ Sgeneric_character_list, 0, 0, 0,
+ "Return a list of all possible generic characters.\n\
+It includes a generic character for a charset not yet defined.")
+ ()
+{
+ return Vgeneric_character_list;
+}
+
+DEFUN ("get-unused-iso-final-char", Fget_unused_iso_final_char,
+ Sget_unused_iso_final_char, 2, 2, 0,
+ "Return an unsed ISO's final char for a charset of DIMENISION and CHARS.\n\
+DIMENSION is the number of bytes to represent a character: 1 or 2.\n\
+CHARS is the number of characters in a dimension: 94 or 96.\n\
+\n\
+This final char is for private use, thus the range is `0' (48) .. `?' (63).\n\
+If there's no unused final char for the specified kind of charset,\n\
+return nil.")
+ (dimension, chars)
+ Lisp_Object dimension, chars;
+{
+ int final_char;
+
+ CHECK_NUMBER (dimension, 0);
+ CHECK_NUMBER (chars, 1);
+ if (XINT (dimension) != 1 && XINT (dimension) != 2)
+ error ("Invalid charset dimension %d, it should be 1 or 2",
+ XINT (dimension));
+ if (XINT (chars) != 94 && XINT (chars) != 96)
+ error ("Invalid charset chars %d, it should be 94 or 96",
+ XINT (chars));
+ for (final_char = '0'; final_char <= '?'; final_char++)
+ {
+ if (ISO_CHARSET_TABLE (dimension, chars, make_number (final_char)) < 0)
+ break;
+ }
+ return (final_char <= '?' ? make_number (final_char) : Qnil);
+}
+
DEFUN ("declare-equiv-charset", Fdeclare_equiv_charset, Sdeclare_equiv_charset,
4, 4, 0,
"Declare a charset of DIMENSION, CHARS, FINAL-CHAR is the same as CHARSET.\n\
/* Return number of different charsets in STR of length LEN. In
addition, for each found charset N, CHARSETS[N] is set 1. The
caller should allocate CHARSETS (MAX_CHARSET + 1 elements) in advance.
- It may lookup a unification table TABLE if supplied. */
+ It may lookup a translation table TABLE if supplied.
+
+ If MULTIBYTE is zero, do not check multibyte characters, i.e. if
+ any ASCII codes (7-bit) are found, CHARSET[0] is set to 1, if any
+ 8-bit codes are found CHARSET[1] is set to 1. */
int
-find_charset_in_str (str, len, charsets, table)
+find_charset_in_str (str, len, charsets, table, multibyte)
unsigned char *str;
int len, *charsets;
Lisp_Object table;
+ int multibyte;
{
- int num = 0;
+ register int num = 0;
+
+ if (! multibyte)
+ {
+ unsigned char *endp = str + len;
+ int maskbits = 0;
+
+ while (str < endp && maskbits != 3)
+ maskbits |= (*str++ < 0x80 ? 1 : 2);
+ if (maskbits & 1)
+ {
+ charsets[0] = 1;
+ num++;
+ }
+ if (maskbits & 2)
+ {
+ charsets[1] = 1;
+ num++;
+ }
+ return num;
+ }
if (! CHAR_TABLE_P (table))
table = Qnil;
while (len > 0)
{
- int bytes = BYTES_BY_CHAR_HEAD (*str);
- int charset;
-
- if (NILP (table))
- charset = CHARSET_AT (str);
- else
- {
- int c, charset;
- unsigned char c1, c2;
+ int bytes, charset, c1, c2;
- SPLIT_STRING(str, bytes, charset, c1, c2);
- if ((c = unify_char (table, -1, charset, c1, c2)) >= 0)
- charset = CHAR_CHARSET (c);
+ SPLIT_MULTIBYTE_SEQ (str, len, bytes, charset, c1, c2);
+
+ if (! NILP (table))
+ {
+ int c1 = translate_char (table, -1, charset, c1, c2);
+ if (c1 >= 0)
+ charset = CHAR_CHARSET (c1);
}
if (!charsets[charset])
2, 3, 0,
"Return a list of charsets in the region between BEG and END.\n\
BEG and END are buffer positions.\n\
-Optional arg TABLE if non-nil is a unification table to look up.")
+Optional arg TABLE if non-nil is a translation table to look up.\n\
+\n\
+If the region contains invalid multiybte characters,\n\
+`unknown' is included in the returned list.\n\
+\n\
+If the current buffer is unibyte, the returned list contains\n\
+`ascii' if any 7-bit characters are found,\n\
+and `unknown' if any 8-bit characters are found.")
(beg, end, table)
Lisp_Object beg, end, table;
{
int charsets[MAX_CHARSET + 1];
- int from, to, stop, i;
+ int from, from_byte, to, stop, stop_byte, i;
Lisp_Object val;
+ int undefined;
+ int multibyte = !NILP (current_buffer->enable_multibyte_characters);
validate_region (&beg, &end);
from = XFASTINT (beg);
stop = to = XFASTINT (end);
+
if (from < GPT && GPT < to)
- stop = GPT;
+ {
+ stop = GPT;
+ stop_byte = GPT_BYTE;
+ }
+ else
+ stop_byte = CHAR_TO_BYTE (stop);
+
+ from_byte = CHAR_TO_BYTE (from);
+
bzero (charsets, (MAX_CHARSET + 1) * sizeof (int));
while (1)
{
- find_charset_in_str (POS_ADDR (from), stop - from, charsets, table);
+ find_charset_in_str (BYTE_POS_ADDR (from_byte), stop_byte - from_byte,
+ charsets, table, multibyte);
if (stop < to)
- from = stop, stop = to;
+ {
+ from = stop, from_byte = stop_byte;
+ stop = to, stop_byte = CHAR_TO_BYTE (stop);
+ }
else
break;
}
+
val = Qnil;
- for (i = MAX_CHARSET; i >= 0; i--)
+ undefined = 0;
+ for (i = (multibyte ? MAX_CHARSET : 1); i >= 0; i--)
if (charsets[i])
- val = Fcons (CHARSET_SYMBOL (i), val);
+ {
+ if (CHARSET_DEFINED_P (i))
+ val = Fcons (CHARSET_SYMBOL (i), val);
+ else
+ undefined = 1;
+ }
+ if (undefined)
+ val = Fcons (Qunknown, val);
return val;
}
DEFUN ("find-charset-string", Ffind_charset_string, Sfind_charset_string,
1, 2, 0,
"Return a list of charsets in STR.\n\
-Optional arg TABLE if non-nil is a unification table to look up.")
+Optional arg TABLE if non-nil is a translation table to look up.\n\
+\n\
+If the region contains invalid multiybte characters,\n\
+`unknown' is included in the returned list.\n\
+\n\
+If STR is unibyte, the returned list contains\n\
+`ascii' if any 7-bit characters are found,\n\
+and `unknown' if any 8-bit characters are found.")
(str, table)
Lisp_Object str, table;
{
int charsets[MAX_CHARSET + 1];
int i;
Lisp_Object val;
+ int undefined;
+ int multibyte;
CHECK_STRING (str, 0);
+ multibyte = STRING_MULTIBYTE (str);
+
bzero (charsets, (MAX_CHARSET + 1) * sizeof (int));
- find_charset_in_str (XSTRING (str)->data, XSTRING (str)->size,
- charsets, table);
+ find_charset_in_str (XSTRING (str)->data, STRING_BYTES (XSTRING (str)),
+ charsets, table, multibyte);
val = Qnil;
- for (i = MAX_CHARSET; i >= 0; i--)
+ undefined = 0;
+ for (i = (multibyte ? MAX_CHARSET : 1); i >= 0; i--)
if (charsets[i])
- val = Fcons (CHARSET_SYMBOL (i), val);
+ {
+ if (CHARSET_DEFINED_P (i))
+ val = Fcons (CHARSET_SYMBOL (i), val);
+ else
+ undefined = 1;
+ }
+ if (undefined)
+ val = Fcons (Qunknown, val);
return val;
}
\f
(charset, code1, code2)
Lisp_Object charset, code1, code2;
{
+ int charset_id, c1, c2;
+
CHECK_NUMBER (charset, 0);
+ charset_id = XINT (charset);
+ if (!CHARSET_DEFINED_P (charset_id))
+ error ("Invalid charset ID: %d", XINT (charset));
if (NILP (code1))
- XSETFASTINT (code1, 0);
+ c1 = 0;
else
- CHECK_NUMBER (code1, 1);
+ {
+ CHECK_NUMBER (code1, 1);
+ c1 = XINT (code1);
+ }
if (NILP (code2))
- XSETFASTINT (code2, 0);
+ c2 = 0;
else
- CHECK_NUMBER (code2, 2);
-
- if (!CHARSET_DEFINED_P (XINT (charset)))
- error ("Invalid charset: %d", XINT (charset));
+ {
+ CHECK_NUMBER (code2, 2);
+ c2 = XINT (code2);
+ }
- return make_number (MAKE_CHAR (XINT (charset), XINT (code1), XINT (code2)));
+ if (c1 < 0 || c1 > 0xFF || c2 < 0 || c2 > 0xFF)
+ error ("Invalid code points for charset ID %d: %d %d", charset_id, c1, c2);
+ c1 &= 0x7F;
+ c2 &= 0x7F;
+ if (c1 == 0
+ ? c2 != 0
+ : (c2 == 0
+ ? !CHAR_COMPONENTS_VALID_P (charset_id, c1, 0x20)
+ : !CHAR_COMPONENTS_VALID_P (charset_id, c1, c2)))
+ error ("Invalid code points for charset ID %d: %d %d", charset_id, c1, c2);
+
+ return make_number (MAKE_CHAR (charset_id, c1, c2));
}
DEFUN ("split-char", Fsplit_char, Ssplit_char, 1, 1, 0,
- "Return list of charset and one or two position-codes of CHAR.")
+ "Return list of charset and one or two position-codes of CHAR.\n\
+If CHAR is invalid as a character code,\n\
+return a list of symbol `unknown' and CHAR.")
(ch)
Lisp_Object ch;
{
Lisp_Object val;
- int charset, c1, c2;
+ int c, charset, c1, c2;
CHECK_NUMBER (ch, 0);
+ c = XFASTINT (ch);
+ if (!CHAR_VALID_P (c, 1))
+ return Fcons (Qunknown, Fcons (ch, Qnil));
SPLIT_CHAR (XFASTINT (ch), charset, c1, c2);
return (c2 >= 0
? Fcons (CHARSET_SYMBOL (charset),
return CHARSET_SYMBOL (CHAR_CHARSET (XINT (ch)));
}
+DEFUN ("charset-after", Fcharset_after, Scharset_after, 0, 1, 0,
+ "Return charset of a character in the current buffer at position POS.\n\
+If POS is nil, it defauls to the current point.\n\
+If POS is out of range, the value is nil.")
+ (pos)
+ Lisp_Object pos;
+{
+ register int pos_byte, bytes, charset, c1, c2;
+ register unsigned char *p;
+
+ if (NILP (pos))
+ pos_byte = PT_BYTE;
+ else if (MARKERP (pos))
+ {
+ pos_byte = marker_byte_position (pos);
+ if (pos_byte < BEGV_BYTE || pos_byte >= ZV_BYTE)
+ return Qnil;
+ }
+ else
+ {
+ CHECK_NUMBER (pos, 0);
+ if (XINT (pos) < BEGV || XINT (pos) >= ZV)
+ return Qnil;
+ pos_byte = CHAR_TO_BYTE (XINT (pos));
+ }
+ p = BYTE_POS_ADDR (pos_byte);
+ if (BASE_LEADING_CODE_P (*p))
+ {
+ SPLIT_MULTIBYTE_SEQ (p, Z_BYTE - pos_byte, bytes, charset, c1, c2);
+ if (charset < 0)
+ charset = 1;
+ }
+ else
+ charset = CHARSET_ASCII;
+
+ return CHARSET_SYMBOL (charset);
+}
+
DEFUN ("iso-charset", Fiso_charset, Siso_charset, 3, 3, 0,
- "Return charset of ISO's specification DIMENSION, CHARS, and FINAL-CHAR.")
+ "Return charset of ISO's specification DIMENSION, CHARS, and FINAL-CHAR.\n\
+\n\
+ISO 2022's designation sequence (escape sequence) distinguishes charsets\n\
+by their DIMENSION, CHARS, and FINAL-CHAR,\n\
+where as Emacs distinguishes them by charset symbol.\n\
+See the documentation of the function `charset-info' for the meanings of\n\
+DIMENSION, CHARS, and FINAL-CHAR.")
(dimension, chars, final_char)
Lisp_Object dimension, chars, final_char;
{
return CHARSET_SYMBOL (charset);
}
+/* If GENERICP is nonzero, return nonzero iff C is a valid normal or
+ generic character. If GENERICP is zero, return nonzero iff C is a
+ valid normal character. Do not call this function directly,
+ instead use macro CHAR_VALID_P. */
+int
+char_valid_p (c, genericp)
+ int c, genericp;
+{
+ int charset, c1, c2;
+
+ if (c < 0)
+ return 0;
+ if (SINGLE_BYTE_CHAR_P (c))
+ return 1;
+ SPLIT_NON_ASCII_CHAR (c, charset, c1, c2);
+ if (genericp)
+ {
+ if (c1)
+ {
+ if (c2 <= 0) c2 = 0x20;
+ }
+ else
+ {
+ if (c2 <= 0) c1 = c2 = 0x20;
+ }
+ }
+ return (CHARSET_DEFINED_P (charset)
+ && CHAR_COMPONENTS_VALID_P (charset, c1, c2));
+}
+
+DEFUN ("char-valid-p", Fchar_valid_p, Schar_valid_p, 1, 2, 0,
+ "Return t if OBJECT is a valid normal character.\n\
+If optional arg GENERICP is non-nil, also return t if OBJECT is\n\
+a valid generic character.")
+ (object, genericp)
+ Lisp_Object object, genericp;
+{
+ if (! NATNUMP (object))
+ return Qnil;
+ return (CHAR_VALID_P (XFASTINT (object), !NILP (genericp)) ? Qt : Qnil);
+}
+
+DEFUN ("unibyte-char-to-multibyte", Funibyte_char_to_multibyte,
+ Sunibyte_char_to_multibyte, 1, 1, 0,
+ "Convert the unibyte character CH to multibyte character.\n\
+The conversion is done based on `nonascii-translation-table' (which see)\n\
+ or `nonascii-insert-offset' (which see).")
+ (ch)
+ Lisp_Object ch;
+{
+ int c;
+
+ CHECK_NUMBER (ch, 0);
+ c = XINT (ch);
+ if (c < 0 || c >= 0400)
+ error ("Invalid unibyte character: %d", c);
+ c = unibyte_char_to_multibyte (c);
+ if (c < 0)
+ error ("Can't convert to multibyte character: %d", XINT (ch));
+ return make_number (c);
+}
+
+DEFUN ("multibyte-char-to-unibyte", Fmultibyte_char_to_unibyte,
+ Smultibyte_char_to_unibyte, 1, 1, 0,
+ "Convert the multibyte character CH to unibyte character.\n\
+The conversion is done based on `nonascii-translation-table' (which see)\n\
+ or `nonascii-insert-offset' (which see).")
+ (ch)
+ Lisp_Object ch;
+{
+ int c;
+
+ CHECK_NUMBER (ch, 0);
+ c = XINT (ch);
+ if (! CHAR_VALID_P (c, 0))
+ error ("Invalid multibyte character: %d", c);
+ c = multibyte_char_to_unibyte (c, Qnil);
+ if (c < 0)
+ error ("Can't convert to unibyte character: %d", XINT (ch));
+ return make_number (c);
+}
+
DEFUN ("char-bytes", Fchar_bytes, Schar_bytes, 1, 1, 0,
- "Return byte length of multi-byte form of CHAR.")
+ "Return 1 regardless of the argument CHAR.\n\
+This is now an obsolete function. We keep it just for backward compatibility.")
(ch)
Lisp_Object ch;
{
Lisp_Object val;
- int bytes;
CHECK_NUMBER (ch, 0);
- if (COMPOSITE_CHAR_P (XFASTINT (ch)))
- {
- unsigned int id = COMPOSITE_CHAR_ID (XFASTINT (ch));
+ return make_number (1);
+}
- bytes = (id < n_cmpchars ? cmpchar_table[id]->len : 1);
- }
- else
- {
- int charset = CHAR_CHARSET (XFASTINT (ch));
+/* Return how many bytes C will occupy in a multibyte buffer.
+ Don't call this function directly, instead use macro CHAR_BYTES. */
+int
+char_bytes (c)
+ int c;
+{
+ int charset;
- bytes = CHARSET_DEFINED_P (charset) ? CHARSET_BYTES (charset) : 1;
- }
+ if (SINGLE_BYTE_CHAR_P (c) || (c & ~((1 << CHARACTERBITS) - 1)))
+ return 1;
- XSETFASTINT (val, bytes);
- return val;
+ charset = CHAR_CHARSET (c);
+ return (CHARSET_DEFINED_P (charset) ? CHARSET_BYTES (charset) : 1);
}
/* Return the width of character of which multi-byte form starts with
: ((! NILP (current_buffer->enable_multibyte_characters) \
&& BASE_LEADING_CODE_P (c)) \
? WIDTH_BY_CHAR_HEAD (c) \
- : 4)))) \
-
+ : 4))))
DEFUN ("char-width", Fchar_width, Schar_width, 1, 1, 0,
"Return width of CHAR when displayed in the current buffer.\n\
XSETINT (val, XVECTOR (disp)->size);
else if (SINGLE_BYTE_CHAR_P (c))
XSETINT (val, ONE_BYTE_CHAR_WIDTH (c));
- else if (COMPOSITE_CHAR_P (c))
- {
- int id = COMPOSITE_CHAR_ID (XFASTINT (ch));
- XSETFASTINT (val, (id < n_cmpchars ? cmpchar_table[id]->width : 0));
- }
else
{
int charset = CHAR_CHARSET (c);
{
unsigned char *endp = str + len;
int width = 0;
- struct Lisp_Char_Table *dp = buffer_display_table (current_buffer);
+ struct Lisp_Char_Table *dp = buffer_display_table ();
while (str < endp)
{
- if (*str == LEADING_CODE_COMPOSITION)
- {
- int id = str_cmpchar_id (str, endp - str);
+ Lisp_Object disp;
+ int thislen;
+ int c = STRING_CHAR_AND_LENGTH (str, endp - str, thislen);
- if (id < 0)
- {
- width += 4;
- str++;
- }
- else
- {
- width += cmpchar_table[id]->width;
- str += cmpchar_table[id]->len;
- }
- }
+ /* Get the way the display table would display it. */
+ if (dp)
+ disp = DISP_CHAR_VECTOR (dp, c);
else
- {
- Lisp_Object disp;
- int thiswidth;
- int c = STRING_CHAR (str, endp - str);
-
- /* Get the way the display table would display it. */
- if (dp)
- disp = DISP_CHAR_VECTOR (dp, c);
- else
- disp = Qnil;
-
- if (VECTORP (disp))
- thiswidth = XVECTOR (disp)->size;
- else
- thiswidth = ONE_BYTE_CHAR_WIDTH (*str);
-
- width += thiswidth;
- str += BYTES_BY_CHAR_HEAD (*str);
- }
+ disp = Qnil;
+
+ if (VECTORP (disp))
+ width += XVECTOR (disp)->size;
+ else
+ width += ONE_BYTE_CHAR_WIDTH (*str);
+
+ str += thislen;
}
return width;
}
DEFUN ("string-width", Fstring_width, Sstring_width, 1, 1, 0,
"Return width of STRING when displayed in the current buffer.\n\
Width is measured by how many columns it occupies on the screen.\n\
-When calculating width of a multi-byte character in STRING,\n\
- only the base leading-code is considered and the validity of\n\
- the following bytes are not checked.")
+When calculating width of a multibyte character in STRING,\n\
+only the base leading-code is considered; the validity of\n\
+the following bytes is not checked.")
(str)
Lisp_Object str;
{
Lisp_Object val;
CHECK_STRING (str, 0);
- XSETFASTINT (val, strwidth (XSTRING (str)->data, XSTRING (str)->size));
+ XSETFASTINT (val, strwidth (XSTRING (str)->data,
+ STRING_BYTES (XSTRING (str))));
return val;
}
CHECK_NUMBER (ch, 0);
charset = CHAR_CHARSET (XFASTINT (ch));
if (!CHARSET_DEFINED_P (charset))
- error ("Invalid character: %d", XINT (ch));
+ invalid_character (XINT (ch));
return CHARSET_TABLE_INFO (charset, CHARSET_DIRECTION_IDX);
}
-DEFUN ("chars-in-string", Fchars_in_string, Schars_in_string, 1, 1, 0,
- "Return number of characters in STRING.")
- (str)
- Lisp_Object str;
+DEFUN ("chars-in-region", Fchars_in_region, Schars_in_region, 2, 2, 0,
+ "Return number of characters between BEG and END.")
+ (beg, end)
+ Lisp_Object beg, end;
{
- Lisp_Object val;
- unsigned char *p, *endp;
- int chars;
+ int from, to;
- CHECK_STRING (str, 0);
+ CHECK_NUMBER_COERCE_MARKER (beg, 0);
+ CHECK_NUMBER_COERCE_MARKER (end, 1);
- p = XSTRING (str)->data; endp = p + XSTRING (str)->size;
- chars = 0;
- while (p < endp)
- {
- if (*p == LEADING_CODE_COMPOSITION)
- {
- p++;
- while (p < endp && ! CHAR_HEAD_P (p)) p++;
- }
- else
- p += BYTES_BY_CHAR_HEAD (*p);
- chars++;
- }
+ from = min (XFASTINT (beg), XFASTINT (end));
+ to = max (XFASTINT (beg), XFASTINT (end));
- XSETFASTINT (val, chars);
- return val;
+ return make_number (to - from);
}
-DEFUN ("char-boundary-p", Fchar_boundary_p, Schar_boundary_p, 1, 1, 0,
- "Return non-nil value if POS is at character boundary of multibyte form.\n\
-The return value is:\n\
- 0 if POS is at an ASCII character or at the end of range,\n\
- 1 if POS is at a head of 2-byte length multi-byte form,\n\
- 2 if POS is at a head of 3-byte length multi-byte form,\n\
- 3 if POS is at a head of 4-byte length multi-byte form,\n\
- 4 if POS is at a head of multi-byte form of a composite character.\n\
-If POS is out of range or not at character boundary, return NIL.")
- (pos)
- Lisp_Object pos;
+/* Return the number of characters in the NBYTES bytes at PTR.
+ This works by looking at the contents and checking for multibyte sequences.
+ However, if the current buffer has enable-multibyte-characters = nil,
+ we treat each byte as a character. */
+
+int
+chars_in_text (ptr, nbytes)
+ unsigned char *ptr;
+ int nbytes;
{
- Lisp_Object val;
- int n;
+ /* current_buffer is null at early stages of Emacs initialization. */
+ if (current_buffer == 0
+ || NILP (current_buffer->enable_multibyte_characters))
+ return nbytes;
- CHECK_NUMBER_COERCE_MARKER (pos, 0);
+ return multibyte_chars_in_text (ptr, nbytes);
+}
- n = XINT (pos);
- if (n < BEGV || n > ZV)
- return Qnil;
+/* Return the number of characters in the NBYTES bytes at PTR.
+ This works by looking at the contents and checking for multibyte sequences.
+ It ignores enable-multibyte-characters. */
- if (n == ZV || NILP (current_buffer->enable_multibyte_characters))
- XSETFASTINT (val, 0);
- else
+int
+multibyte_chars_in_text (ptr, nbytes)
+ unsigned char *ptr;
+ int nbytes;
+{
+ unsigned char *endp;
+ int chars, bytes;
+
+ endp = ptr + nbytes;
+ chars = 0;
+
+ while (ptr < endp)
{
- unsigned char *p = POS_ADDR (n);
-
- if (SINGLE_BYTE_CHAR_P (*p))
- XSETFASTINT (val, 0);
- else if (*p == LEADING_CODE_COMPOSITION)
- XSETFASTINT (val, 4);
- else if (BYTES_BY_CHAR_HEAD (*p) > 1)
- XSETFASTINT (val, BYTES_BY_CHAR_HEAD (*p) - 1);
+ if (BASE_LEADING_CODE_P (*ptr))
+ {
+ PARSE_MULTIBYTE_SEQ (ptr, nbytes, bytes);
+ ptr += bytes;
+ nbytes -= bytes;
+ }
else
- val = Qnil;
+ {
+ ptr++;
+ nbytes--;
+ }
+ chars++;
}
- return val;
+
+ return chars;
}
-DEFUN ("concat-chars", Fconcat_chars, Sconcat_chars, 1, MANY, 0,
+DEFUN ("string", Fstring, Sstring, 1, MANY, 0,
"Concatenate all the argument characters and make the result a string.")
(n, args)
int n;
Lisp_Object *args;
{
int i;
- unsigned char *buf
- = (unsigned char *) alloca (MAX_LENGTH_OF_MULTI_BYTE_FORM * n);
+ unsigned char *buf = (unsigned char *) alloca (MAX_MULTIBYTE_LENGTH * n);
unsigned char *p = buf;
Lisp_Object val;
+ int c, multibyte_p = 0;
for (i = 0; i < n; i++)
{
- int c, len;
- unsigned char *str;
-
- if (!INTEGERP (args[i]))
- {
- free (buf);
- CHECK_NUMBER (args[i], 0);
- }
+ CHECK_NUMBER (args[i], 0);
c = XINT (args[i]);
- len = CHAR_STRING (c, p, str);
- if (p != str)
- /* C is a composite character. */
- bcopy (str, p, len);
- p += len;
+ p += CHAR_STRING (c, p);
+
+ if (!SINGLE_BYTE_CHAR_P (c))
+ multibyte_p = 1;
}
- val = make_string (buf, p - buf);
+ /* Here, we can't use make_string_from_bytes because of the byte
+ combining problem. Make a multibyte string if there is any
+ multibyte character in ARGS to make sure that `(string 2276)'
+ returns a multibyte string if running --unibyte. */
+ if (multibyte_p)
+ val = make_multibyte_string (buf, n, p - buf);
+ else
+ val = make_unibyte_string (buf, p - buf);
return val;
}
#endif /* emacs */
\f
-/*** Composite characters staffs ***/
-
-/* Each composite character is identified by CMPCHAR-ID which is
- assigned when Emacs needs the character code of the composite
- character (e.g. when displaying it on the screen). See the
- document "GENERAL NOTE on COMPOSITE CHARACTER" in `charset.h' how a
- composite character is represented in Emacs. */
-
-/* If `static' is defined, it means that it is defined to null string. */
-#ifndef static
-/* The following function is copied from lread.c. */
-static int
-hash_string (ptr, len)
- unsigned char *ptr;
- int len;
-{
- register unsigned char *p = ptr;
- register unsigned char *end = p + len;
- register unsigned char c;
- register int hash = 0;
-
- while (p != end)
- {
- c = *p++;
- if (c >= 0140) c -= 40;
- hash = ((hash<<3) + (hash>>28) + c);
- }
- return hash & 07777777777;
-}
-#endif
-
-#define CMPCHAR_HASH_TABLE_SIZE 0xFFF
-
-static int *cmpchar_hash_table[CMPCHAR_HASH_TABLE_SIZE];
-
-/* Each element of `cmpchar_hash_table' is a pointer to an array of
- integer, where the 1st element is the size of the array, the 2nd
- element is how many elements are actually used in the array, and
- the remaining elements are CMPCHAR-IDs of composite characters of
- the same hash value. */
-#define CMPCHAR_HASH_SIZE(table) table[0]
-#define CMPCHAR_HASH_USED(table) table[1]
-#define CMPCHAR_HASH_CMPCHAR_ID(table, i) table[i]
-
-/* Return CMPCHAR-ID of the composite character in STR of the length
- LEN. If the composite character has not yet been registered,
- register it in `cmpchar_table' and assign new CMPCHAR-ID. This
- is the sole function for assigning CMPCHAR-ID. */
-int
-str_cmpchar_id (str, len)
- unsigned char *str;
- int len;
-{
- int hash_idx, *hashp;
- unsigned char *buf;
- int embedded_rule; /* 1 if composition rule is embedded. */
- int chars; /* number of components. */
- int i;
- struct cmpchar_info *cmpcharp;
-
- if (len < 5)
- /* Any composite char have at least 3-byte length. */
- return -1;
-
- /* The second byte 0xFF means compostion rule is embedded. */
- embedded_rule = (str[1] == 0xFF);
-
- /* At first, get the actual length of the composite character. */
- {
- unsigned char *p, *endp = str + 1, *lastp = str + len;
- int bytes;
-
- while (endp < lastp && ! CHAR_HEAD_P (endp)) endp++;
- chars = 0;
- p = str + 1 + embedded_rule;
- while (p < endp)
- {
- /* No need of checking if *P is 0xA0 because
- BYTES_BY_CHAR_HEAD (0x80) surely returns 2. */
- p += (bytes = BYTES_BY_CHAR_HEAD (*p - 0x20) + embedded_rule);
- chars++;
- }
- len = (p -= embedded_rule) - str;
- if (p > endp)
- len -= - bytes, chars--;
-
- if (chars < 2 || chars > MAX_COMPONENT_COUNT)
- /* Invalid number of components. */
- return -1;
- }
- hash_idx = hash_string (str, len) % CMPCHAR_HASH_TABLE_SIZE;
- hashp = cmpchar_hash_table[hash_idx];
-
- /* Then, look into the hash table. */
- if (hashp != NULL)
- /* Find the correct one among composite characters of the same
- hash value. */
- for (i = 2; i < CMPCHAR_HASH_USED (hashp); i++)
- {
- cmpcharp = cmpchar_table[CMPCHAR_HASH_CMPCHAR_ID (hashp, i)];
- if (len == cmpcharp->len
- && ! bcmp (str, cmpcharp->data, len))
- return CMPCHAR_HASH_CMPCHAR_ID (hashp, i);
- }
-
- /* We have to register the composite character in cmpchar_table. */
- if (n_cmpchars > (CHAR_FIELD2_MASK | CHAR_FIELD3_MASK))
- /* No, we have no more room for a new composite character. */
- return -1;
-
- /* Make the entry in hash table. */
- if (hashp == NULL)
- {
- /* Make a table for 8 composite characters initially. */
- hashp = (cmpchar_hash_table[hash_idx]
- = (int *) xmalloc (sizeof (int) * (2 + 8)));
- CMPCHAR_HASH_SIZE (hashp) = 10;
- CMPCHAR_HASH_USED (hashp) = 2;
- }
- else if (CMPCHAR_HASH_USED (hashp) >= CMPCHAR_HASH_SIZE (hashp))
- {
- CMPCHAR_HASH_SIZE (hashp) += 8;
- hashp = (cmpchar_hash_table[hash_idx]
- = (int *) xrealloc (hashp,
- sizeof (int) * CMPCHAR_HASH_SIZE (hashp)));
- }
- CMPCHAR_HASH_CMPCHAR_ID (hashp, CMPCHAR_HASH_USED (hashp)) = n_cmpchars;
- CMPCHAR_HASH_USED (hashp)++;
-
- /* Set information of the composite character in cmpchar_table. */
- if (cmpchar_table_size == 0)
- {
- /* This is the first composite character to be registered. */
- cmpchar_table_size = 256;
- cmpchar_table
- = (struct cmpchar_info **) xmalloc (sizeof (cmpchar_table[0])
- * cmpchar_table_size);
- }
- else if (cmpchar_table_size <= n_cmpchars)
- {
- cmpchar_table_size += 256;
- cmpchar_table
- = (struct cmpchar_info **) xrealloc (cmpchar_table,
- sizeof (cmpchar_table[0])
- * cmpchar_table_size);
- }
-
- cmpcharp = (struct cmpchar_info *) xmalloc (sizeof (struct cmpchar_info));
-
- cmpcharp->len = len;
- cmpcharp->data = (unsigned char *) xmalloc (len + 1);
- bcopy (str, cmpcharp->data, len);
- cmpcharp->data[len] = 0;
- cmpcharp->glyph_len = chars;
- cmpcharp->glyph = (GLYPH *) xmalloc (sizeof (GLYPH) * chars);
- if (embedded_rule)
- {
- cmpcharp->cmp_rule = (unsigned char *) xmalloc (chars);
- cmpcharp->col_offset = (float *) xmalloc (sizeof (float) * chars);
- }
- else
- {
- cmpcharp->cmp_rule = NULL;
- cmpcharp->col_offset = NULL;
- }
-
- /* Setup GLYPH data and composition rules (if any) so as not to make
- them every time on displaying. */
- {
- unsigned char *bufp;
- int width;
- float leftmost = 0.0, rightmost = 1.0;
-
- if (embedded_rule)
- /* At first, col_offset[N] is set to relative to col_offset[0]. */
- cmpcharp->col_offset[0] = 0;
-
- for (i = 0, bufp = cmpcharp->data + 1; i < chars; i++)
- {
- if (embedded_rule)
- cmpcharp->cmp_rule[i] = *bufp++;
-
- if (*bufp == 0xA0) /* This is an ASCII character. */
- {
- cmpcharp->glyph[i] = FAST_MAKE_GLYPH ((*++bufp & 0x7F), 0);
- width = 1;
- bufp++;
- }
- else /* Multibyte character. */
- {
- /* Make `bufp' point normal multi-byte form temporally. */
- *bufp -= 0x20;
- cmpcharp->glyph[i]
- = FAST_MAKE_GLYPH (string_to_non_ascii_char (bufp, 4, 0), 0);
- width = WIDTH_BY_CHAR_HEAD (*bufp);
- *bufp += 0x20;
- bufp += BYTES_BY_CHAR_HEAD (*bufp - 0x20);
- }
-
- if (embedded_rule && i > 0)
- {
- /* Reference points (global_ref and new_ref) are
- encoded as below:
-
- 0--1--2 -- ascent
- | |
- | |
- | 4 -+--- center
- -- 3 5 -- baseline
- | |
- 6--7--8 -- descent
-
- Now, we calculate the column offset of the new glyph
- from the left edge of the first glyph. This can avoid
- the same calculation everytime displaying this
- composite character. */
-
- /* Reference points of global glyph and new glyph. */
- int global_ref = (cmpcharp->cmp_rule[i] - 0xA0) / 9;
- int new_ref = (cmpcharp->cmp_rule[i] - 0xA0) % 9;
- /* Column offset relative to the first glyph. */
- float left = (leftmost
- + (global_ref % 3) * (rightmost - leftmost) / 2.0
- - (new_ref % 3) * width / 2.0);
-
- cmpcharp->col_offset[i] = left;
- if (left < leftmost)
- leftmost = left;
- if (left + width > rightmost)
- rightmost = left + width;
- }
- else
- {
- if (width > rightmost)
- rightmost = width;
- }
- }
- if (embedded_rule)
- {
- /* Now col_offset[N] are relative to the left edge of the
- first component. Make them relative to the left edge of
- overall glyph. */
- for (i = 0; i < chars; i++)
- cmpcharp->col_offset[i] -= leftmost;
- /* Make rightmost holds width of overall glyph. */
- rightmost -= leftmost;
- }
-
- cmpcharp->width = rightmost;
- if (cmpcharp->width < rightmost)
- /* To get a ceiling integer value. */
- cmpcharp->width++;
- }
-
- cmpchar_table[n_cmpchars] = cmpcharp;
-
- return n_cmpchars++;
-}
-
-/* Return the Nth element of the composite character C. */
int
-cmpchar_component (c, n)
- unsigned int c, n;
-{
- int id = COMPOSITE_CHAR_ID (c);
-
- if (id >= n_cmpchars /* C is not a valid composite character. */
- || n >= cmpchar_table[id]->glyph_len) /* No such component. */
- return -1;
- /* No face data is stored in glyph code. */
- return ((int) (cmpchar_table[id]->glyph[n]));
-}
-
-DEFUN ("cmpcharp", Fcmpcharp, Scmpcharp, 1, 1, 0,
- "T if CHAR is a composite character.")
- (ch)
- Lisp_Object ch;
-{
- CHECK_NUMBER (ch, 0);
- return (COMPOSITE_CHAR_P (XINT (ch)) ? Qt : Qnil);
-}
-
-DEFUN ("composite-char-component", Fcmpchar_component, Scmpchar_component,
- 2, 2, 0,
- "Return the IDXth component character of composite character CHARACTER.")
- (character, idx)
- Lisp_Object character, idx;
-{
- int c;
-
- CHECK_NUMBER (character, 0);
- CHECK_NUMBER (idx, 1);
-
- if ((c = cmpchar_component (XINT (character), XINT (idx))) < 0)
- args_out_of_range (character, idx);
-
- return make_number (c);
-}
-
-DEFUN ("composite-char-composition-rule", Fcmpchar_cmp_rule, Scmpchar_cmp_rule,
- 2, 2, 0,
- "Return the Nth composition rule embedded in composite character CHARACTER.\n\
-The returned rule is for composing the Nth component\n\
-on the (N-1)th component. If N is 0, the returned value is always 255.")
- (character, n)
- Lisp_Object character, n;
-{
- int id, i;
-
- CHECK_NUMBER (character, 0);
- CHECK_NUMBER (n, 1);
-
- id = COMPOSITE_CHAR_ID (XINT (character));
- if (id < 0 || id >= n_cmpchars)
- error ("Invalid composite character: %d", XINT (character));
- i = XINT (n);
- if (i > cmpchar_table[id]->glyph_len)
- args_out_of_range (character, n);
-
- return make_number (cmpchar_table[id]->cmp_rule[i]);
-}
-
-DEFUN ("composite-char-composition-rule-p", Fcmpchar_cmp_rule_p,
- Scmpchar_cmp_rule_p, 1, 1, 0,
- "Return non-nil if composite character CHARACTER contains a embedded rule.")
- (character)
- Lisp_Object character;
-{
- int id;
-
- CHECK_NUMBER (character, 0);
- id = COMPOSITE_CHAR_ID (XINT (character));
- if (id < 0 || id >= n_cmpchars)
- error ("Invalid composite character: %d", XINT (character));
-
- return (cmpchar_table[id]->cmp_rule ? Qt : Qnil);
-}
-
-DEFUN ("composite-char-component-count", Fcmpchar_cmp_count,
- Scmpchar_cmp_count, 1, 1, 0,
- "Return number of compoents of composite character CHARACTER.")
- (character)
- Lisp_Object character;
-{
- int id;
-
- CHECK_NUMBER (character, 0);
- id = COMPOSITE_CHAR_ID (XINT (character));
- if (id < 0 || id >= n_cmpchars)
- error ("Invalid composite character: %d", XINT (character));
-
- return (make_number (cmpchar_table[id]->glyph_len));
-}
-
-DEFUN ("compose-string", Fcompose_string, Scompose_string,
- 1, 1, 0,
- "Return one char string composed from all characters in STRING.")
- (str)
- Lisp_Object str;
-{
- unsigned char buf[MAX_LENGTH_OF_MULTI_BYTE_FORM], *p, *pend, *ptemp;
- int len, i;
-
- CHECK_STRING (str, 0);
-
- buf[0] = LEADING_CODE_COMPOSITION;
- p = XSTRING (str)->data;
- pend = p + XSTRING (str)->size;
- i = 1;
- while (p < pend)
- {
- if (*p < 0x20 || *p == 127) /* control code */
- error ("Invalid component character: %d", *p);
- else if (*p < 0x80) /* ASCII */
- {
- if (i + 2 >= MAX_LENGTH_OF_MULTI_BYTE_FORM)
- error ("Too long string to be composed: %s", XSTRING (str)->data);
- /* Prepend an ASCII charset indicator 0xA0, set MSB of the
- code itself. */
- buf[i++] = 0xA0;
- buf[i++] = *p++ + 0x80;
- }
- else if (*p == LEADING_CODE_COMPOSITION) /* composite char */
- {
- /* Already composed. Eliminate the heading
- LEADING_CODE_COMPOSITION, keep the remaining bytes
- unchanged. */
- p++;
- ptemp = p;
- while (! CHAR_HEAD_P (p)) p++;
- if (i + (p - ptemp) >= MAX_LENGTH_OF_MULTI_BYTE_FORM)
- error ("Too long string to be composed: %s", XSTRING (str)->data);
- bcopy (ptemp, buf + i, p - ptemp);
- i += p - ptemp;
- }
- else /* multibyte char */
- {
- /* Add 0x20 to the base leading-code, keep the remaining
- bytes unchanged. */
- len = BYTES_BY_CHAR_HEAD (*p);
- if (i + len >= MAX_LENGTH_OF_MULTI_BYTE_FORM)
- error ("Too long string to be composed: %s", XSTRING (str)->data);
- bcopy (p, buf + i, len);
- buf[i] += 0x20;
- p += len, i += len;
- }
- }
-
- if (i < 5)
- /* STR contains only one character, which can't be composed. */
- error ("Too short string to be composed: %s", XSTRING (str)->data);
-
- return make_string (buf, i);
-}
-
-\f
charset_id_internal (charset_name)
char *charset_name;
{
- Lisp_Object val = Fget (intern (charset_name), Qcharset);
+ Lisp_Object val;
+ val= Fget (intern (charset_name), Qcharset);
if (!VECTORP (val))
error ("Charset %s is not defined", charset_name);
return Qnil;
}
+void
init_charset_once ()
{
int i, j, k;
staticpro (&Vcharset_table);
staticpro (&Vcharset_symbol_table);
+ staticpro (&Vgeneric_character_list);
/* This has to be done here, before we call Fmake_char_table. */
Qcharset_table = intern ("charset-table");
Fput (Qcharset_table, Qchar_table_extra_slots, make_number (0));
Vcharset_table = Fmake_char_table (Qcharset_table, Qnil);
- Vcharset_symbol_table = Fmake_vector (make_number (MAX_CHARSET + 1), Qnil);
+ Qunknown = intern ("unknown");
+ staticpro (&Qunknown);
+ Vcharset_symbol_table = Fmake_vector (make_number (MAX_CHARSET + 1),
+ Qunknown);
/* Setup tables. */
for (i = 0; i < 2; i++)
for (k = 0; k < 128; k++)
iso_charset_table [i][j][k] = -1;
- bzero (cmpchar_hash_table, sizeof cmpchar_hash_table);
- cmpchar_table_size = n_cmpchars = 0;
-
for (i = 0; i < 256; i++)
BYTES_BY_CHAR_HEAD (i) = 1;
+ for (i = MIN_CHARSET_OFFICIAL_DIMENSION1;
+ i <= MAX_CHARSET_OFFICIAL_DIMENSION1; i++)
+ BYTES_BY_CHAR_HEAD (i) = 2;
+ for (i = MIN_CHARSET_OFFICIAL_DIMENSION2;
+ i <= MAX_CHARSET_OFFICIAL_DIMENSION2; i++)
+ BYTES_BY_CHAR_HEAD (i) = 3;
BYTES_BY_CHAR_HEAD (LEADING_CODE_PRIVATE_11) = 3;
BYTES_BY_CHAR_HEAD (LEADING_CODE_PRIVATE_12) = 3;
BYTES_BY_CHAR_HEAD (LEADING_CODE_PRIVATE_21) = 4;
BYTES_BY_CHAR_HEAD (LEADING_CODE_PRIVATE_22) = 4;
- /* The following doesn't reflect the actual bytes, but just to tell
+ /* The followings don't reflect the actual bytes, but just to tell
that it is a start of a multibyte character. */
- BYTES_BY_CHAR_HEAD (LEADING_CODE_COMPOSITION) = 2;
+ BYTES_BY_CHAR_HEAD (0x80) = 2;
+ BYTES_BY_CHAR_HEAD (0x9E) = 2;
+ BYTES_BY_CHAR_HEAD (0x9F) = 2;
for (i = 0; i < 128; i++)
WIDTH_BY_CHAR_HEAD (i) = 1;
WIDTH_BY_CHAR_HEAD (LEADING_CODE_PRIVATE_12) = 2;
WIDTH_BY_CHAR_HEAD (LEADING_CODE_PRIVATE_21) = 1;
WIDTH_BY_CHAR_HEAD (LEADING_CODE_PRIVATE_22) = 2;
+
+ {
+ Lisp_Object val;
+
+ val = Qnil;
+ for (i = 0x81; i < 0x90; i++)
+ val = Fcons (make_number ((i - 0x70) << 7), val);
+ for (; i < 0x9A; i++)
+ val = Fcons (make_number ((i - 0x8F) << 14), val);
+ for (i = 0xA0; i < 0xF0; i++)
+ val = Fcons (make_number ((i - 0x70) << 7), val);
+ for (; i < 0xFF; i++)
+ val = Fcons (make_number ((i - 0xE0) << 14), val);
+ Vgeneric_character_list = Fnreverse (val);
+ }
+
+ nonascii_insert_offset = 0;
+ Vnonascii_translation_table = Qnil;
}
#ifdef emacs
+void
syms_of_charset ()
{
Qascii = intern ("ascii");
CHARSET_SYMBOL (CHARSET_ASCII) = Qascii;
Fput (Qascii, Qcharset, CHARSET_TABLE_ENTRY (CHARSET_ASCII));
- Qcomposition = intern ("composition");
- staticpro (&Qcomposition);
- CHARSET_SYMBOL (CHARSET_COMPOSITION) = Qcomposition;
+ Qauto_fill_chars = intern ("auto-fill-chars");
+ staticpro (&Qauto_fill_chars);
+ Fput (Qauto_fill_chars, Qchar_table_extra_slots, make_number (0));
defsubr (&Sdefine_charset);
+ defsubr (&Sgeneric_character_list);
+ defsubr (&Sget_unused_iso_final_char);
defsubr (&Sdeclare_equiv_charset);
defsubr (&Sfind_charset_region);
defsubr (&Sfind_charset_string);
defsubr (&Smake_char_internal);
defsubr (&Ssplit_char);
defsubr (&Schar_charset);
+ defsubr (&Scharset_after);
defsubr (&Siso_charset);
+ defsubr (&Schar_valid_p);
+ defsubr (&Sunibyte_char_to_multibyte);
+ defsubr (&Smultibyte_char_to_unibyte);
defsubr (&Schar_bytes);
defsubr (&Schar_width);
defsubr (&Sstring_width);
defsubr (&Schar_direction);
- defsubr (&Schars_in_string);
- defsubr (&Schar_boundary_p);
- defsubr (&Sconcat_chars);
- defsubr (&Scmpcharp);
- defsubr (&Scmpchar_component);
- defsubr (&Scmpchar_cmp_rule);
- defsubr (&Scmpchar_cmp_rule_p);
- defsubr (&Scmpchar_cmp_count);
- defsubr (&Scompose_string);
+ defsubr (&Schars_in_region);
+ defsubr (&Sstring);
defsubr (&Ssetup_special_charsets);
DEFVAR_LISP ("charset-list", &Vcharset_list,
"List of charsets ever defined.");
Vcharset_list = Fcons (Qascii, Qnil);
- DEFVAR_INT ("leading-code-composition", &leading_code_composition,
- "Leading-code of composite characters.");
- leading_code_composition = LEADING_CODE_COMPOSITION;
+ DEFVAR_LISP ("translation-table-vector", &Vtranslation_table_vector,
+ "Vector of cons cell of a symbol and translation table ever defined.\n\
+An ID of a translation table is an index of this vector.");
+ Vtranslation_table_vector = Fmake_vector (make_number (16), Qnil);
DEFVAR_INT ("leading-code-private-11", &leading_code_private_11,
"Leading-code of private TYPE9N charset of column-width 1.");
DEFVAR_INT ("leading-code-private-22", &leading_code_private_22,
"Leading-code of private TYPE9Nx9N charset of column-width 2.");
leading_code_private_22 = LEADING_CODE_PRIVATE_22;
+
+ DEFVAR_INT ("nonascii-insert-offset", &nonascii_insert_offset,
+ "Offset for converting non-ASCII unibyte codes 0240...0377 to multibyte.\n\
+This is used for converting unibyte text to multibyte,\n\
+and for inserting character codes specified by number.\n\n\
+This serves to convert a Latin-1 or similar 8-bit character code\n\
+to the corresponding Emacs multibyte character code.\n\
+Typically the value should be (- (make-char CHARSET 0) 128),\n\
+for your choice of character set.\n\
+If `nonascii-translation-table' is non-nil, it overrides this variable.");
+ nonascii_insert_offset = 0;
+
+ DEFVAR_LISP ("nonascii-translation-table", &Vnonascii_translation_table,
+ "Translation table to convert non-ASCII unibyte codes to multibyte.\n\
+This is used for converting unibyte text to multibyte,\n\
+and for inserting character codes specified by number.\n\n\
+Conversion is performed only when multibyte characters are enabled,\n\
+and it serves to convert a Latin-1 or similar 8-bit character code\n\
+to the corresponding Emacs character code.\n\n\
+If this is nil, `nonascii-insert-offset' is used instead.\n\
+See also the docstring of `make-translation-table'.");
+ Vnonascii_translation_table = Qnil;
+
+ DEFVAR_LISP ("auto-fill-chars", &Vauto_fill_chars,
+ "A char-table for characters which invoke auto-filling.\n\
+Such characters has value t in this table.");
+ Vauto_fill_chars = Fmake_char_table (Qauto_fill_chars, Qnil);
+ CHAR_TABLE_SET (Vauto_fill_chars, make_number (' '), Qt);
+ CHAR_TABLE_SET (Vauto_fill_chars, make_number ('\n'), Qt);
}
#endif /* emacs */