#endif /* emacs */
Lisp_Object Qcharset, Qascii, Qcomposition;
+Lisp_Object Qunknown;
/* Declaration of special leading-codes. */
int leading_code_composition; /* for composite characters */
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];
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
invalid_character (c)
int c;
{
- error ("Invalid character: %o, %d, 0x%x", c);
+ error ("Invalid character: 0%o, %d, 0x%x", c, c, c);
}
+/* Parse composite character string STR of length LENGTH (>= 2) and
+ set BYTES, CHARSET, C1, and C2 as below.
+
+ It is assumed that *STR is LEADING_CODE_COMPOSITION and the
+ following (LENGTH - 1) bytes satisfy !CHAR_HEAD_P.
+
+ If there is a valid composite character, set CHARSET, C1, and C2 to
+ such values that MAKE_CHAR can make the composite character from
+ them. Otherwise, set CHARSET to CHARSET_COMPOSITION, set C1 to the
+ second byte of the sequence, C2 to -1 so that MAKE_CHAR can make
+ the invalid multibyte character whose string representation is two
+ bytes of STR[0] and STR[1]. In any case, set BYTES to LENGTH.
+
+ This macro should be called only from SPLIT_MULTIBYTE_SEQ. */
+
+#define SPLIT_COMPOSITE_SEQ(str, length, bytes, charset, c1, c2) \
+ do { \
+ int cmpchar_id = str_cmpchar_id ((str), (length)); \
+ \
+ (charset) = CHARSET_COMPOSITION; \
+ (bytes) = (length); \
+ if (cmpchar_id >= 0) \
+ { \
+ (c1) = CHAR_FIELD2 (cmpchar_id); \
+ (c2) = CHAR_FIELD3 (cmpchar_id); \
+ } \
+ else \
+ { \
+ (c1) = (str)[1] & 0x7F; \
+ (c2) = -1; \
+ } \
+ } while (0)
+
+/* Parse non-composite multibyte character string STR of length LENGTH
+ (>= 2) and 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 (excluding
+ LEADING_CODE_COMPOSITION) 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 ((str)[0] == LEADING_CODE_COMPOSITION) \
+ SPLIT_COMPOSITE_SEQ (str, i, bytes, charset, c1, c2); \
+ 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_DIMENSION (charset) == 1 \
+ ? ((c1) >= 0x20 && (c1) <= 0x7F) \
+ : ((c1) >= 0x20 && (c1) <= 0x7F && (c2) >= 0x20 && (c2) <= 0x7F))
/* 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,
- signal an error.
+ form. If C is an invalid character, store (C & 0xFF) in WORKBUF[0]
+ and return 1.
Use macro `CHAR_STRING (C, WORKBUF, STR)' instead of calling this
function directly if C can be an ASCII character. */
int c;
unsigned char *workbuf, **str;
{
- int charset, c1, c2;
-
- 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)
{
- invalid_character (c);
+ /* 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);
+
+ *str = workbuf;
+ *workbuf++ = c;
}
+ else
+ {
+ int charset, c1, c2;
- SPLIT_NON_ASCII_CHAR (c, charset, c1, c2);
- if (!charset
- || ! CHARSET_DEFINED_P (charset)
- || c1 >= 0 && c1 < 32
- || c2 >= 0 && c2 < 32)
- invalid_character (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;
+ SPLIT_NON_ASCII_CHAR (c, charset, c1, c2);
+ if (charset == CHARSET_COMPOSITION)
+ {
+ if (c >= MAX_CHAR)
+ invalid_character (c);
+ if (c >= MIN_CHAR_COMPOSITION)
+ {
+ /* Valid composite character. */
+ *str = cmpchar_table[COMPOSITE_CHAR_ID (c)]->data;
+ workbuf = *str + cmpchar_table[COMPOSITE_CHAR_ID (c)]->len;
+ }
+ else
+ {
+ /* Invalid but can have multibyte form. */
+ *str = workbuf;
+ *workbuf++ = LEADING_CODE_COMPOSITION;
+ *workbuf++ = c1 | 0x80;
+ }
+ }
+ else if (charset > CHARSET_COMPOSITION)
+ {
+ *str = workbuf;
+ if (charset >= LEADING_CODE_EXT_11)
+ *workbuf++ = (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)));
+ *workbuf++ = charset;
+ if (c1 > 0 && c1 < 32 || c2 > 0 && c2 < 32)
+ invalid_character (c);
+ if (c1)
+ {
+ *workbuf++ = c1 | 0x80;
+ if (c2 > 0)
+ *workbuf++ = c2 | 0x80;
+ }
+ }
+ else if (charset == CHARSET_ASCII)
+ *workbuf++= c & 0x7F;
+ else
+ invalid_character (c);
+ }
return (workbuf - *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
- multibyte form is set to the address ACTUAL_LEN.
-
- If exclude_tail_garbage is nonzero, ACTUAL_LEN excludes gabage
- bytes following the non-ASCII character.
+/* 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. */
+ directly if you want ot handle ASCII characters as well. */
int
-string_to_non_ascii_char (str, len, actual_len, exclude_tail_garbage)
+string_to_non_ascii_char (str, len, actual_len)
const unsigned char *str;
- int len, *actual_len, exclude_tail_garbage;
+ int len, *actual_len;
{
- int charset;
- unsigned char c1, c2;
- register int c, bytes;
-
- c = *str;
- bytes = 1;
-
- if (BASE_LEADING_CODE_P (c))
- {
- while (bytes < len && ! CHAR_HEAD_P (str[bytes])) bytes++;
-
- if (c == LEADING_CODE_COMPOSITION)
- {
- int cmpchar_id = str_cmpchar_id (str, bytes);
-
- if (cmpchar_id >= 0)
- c = MAKE_COMPOSITE_CHAR (cmpchar_id);
- if (exclude_tail_garbage)
- bytes = cmpchar_table[cmpchar_id]->len;
- }
- else
- {
- int charset = c, c1, c2 = 0;
- int char_bytes = BYTES_BY_CHAR_HEAD (c);
-
- str++;
- if (c >= LEADING_CODE_PRIVATE_11)
- charset = *str++;
- if (char_bytes <= bytes && CHARSET_DEFINED_P (charset))
- {
- c1 = *str++ & 0x7f;
- if (CHARSET_DIMENSION (charset) == 2)
- c2 = *str & 0x7F;
- c = MAKE_NON_ASCII_CHAR (charset, c1, c2);
- if (exclude_tail_garbage)
- bytes = char_bytes;
- }
- }
- }
+ 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 = 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)
const unsigned char *str;
int len;
{
- int bytes = 1;
-
- if (BASE_LEADING_CODE_P (*str))
- while (bytes < len && ! CHAR_HEAD_P (str[bytes])) bytes++;
+ int bytes;
+ 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. */
int
split_non_ascii_string (str, len, charset, c1, c2)
- register const unsigned char *str;
- register unsigned char *c1, *c2;
- register int len, *charset;
+ 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;
- return 0;
+ *charset = cs;
+ *c1 = code1;
+ *c2 = code2;
+}
+
+/* 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 >= MIN_CHAR_COMPOSITION)
+ return (c < MAX_CHAR);
+
+ 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
if (c < 0) c = MAKE_CHAR (charset, c1, c2);
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);
unibyte_char_to_multibyte (c)
int c;
{
- if (c >= 0240 && c < 0400)
+ if (c < 0400 && c >= 0200)
{
int c_save = c;
if (! NILP (Vnonascii_translation_table))
- c = XINT (Faref (Vnonascii_translation_table, make_number (c)));
- else if (nonascii_insert_offset > 0)
- c += nonascii_insert_offset;
- if (c >= 0240 && (c < 0400 || ! VALID_MULTIBYTE_CHAR_P (c)))
+ {
+ 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).
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. */
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\
If CMPCHARP is nonzero and some composite character is found,
CHARSETS[128] is also set 1 and the returned number is incremented
- by 1. */
+ by 1.
+
+ 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, cmpcharp)
+find_charset_in_str (str, len, charsets, table, cmpcharp, multibyte)
unsigned char *str;
int len, *charsets;
Lisp_Object table;
int cmpcharp;
+ int multibyte;
{
register int num = 0, c;
+ 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;
if (cmpchar_id >= 0)
{
- struct cmpchar_info *cmpcharp = cmpchar_table[cmpchar_id];
+ struct cmpchar_info *cmp_p = cmpchar_table[cmpchar_id];
int i;
- for (i = 0; i < cmpcharp->glyph_len; i++)
+ for (i = 0; i < cmp_p->glyph_len; i++)
{
- c = cmpcharp->glyph[i];
+ c = cmp_p->glyph[i];
if (!NILP (table))
{
if ((c = translate_char (table, c, 0, 0, 0)) < 0)
- c = cmpcharp->glyph[i];
+ c = cmp_p->glyph[i];
}
if ((charset = CHAR_CHARSET (c)) < 0)
charset = CHARSET_ASCII;
num += 1;
}
}
- str += cmpcharp->len;
- len -= cmpcharp->len;
- if (!charsets[LEADING_CODE_COMPOSITION])
+ str += cmp_p->len;
+ len -= cmp_p->len;
+ if (cmpcharp && !charsets[CHARSET_COMPOSITION])
{
- charsets[LEADING_CODE_COMPOSITION] = 1;
+ charsets[CHARSET_COMPOSITION] = 1;
num += 1;
}
continue;
}
- charset = CHARSET_ASCII;
+ charset = 1; /* This leads to `unknown' charset. */
bytes = 1;
}
else
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 translation table to look up.")
+If the region contains any composite character,\n\
+`composition' is included in the returned list.\n\
+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, 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);
while (1)
{
find_charset_in_str (BYTE_POS_ADDR (from_byte), stop_byte - from_byte,
- charsets, table, 0);
+ charsets, table, 1, multibyte);
if (stop < to)
{
from = stop, from_byte = stop_byte;
}
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) || i == CHARSET_COMPOSITION)
+ 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 translation table to look up.")
+If the string contains any composite characters,\n\
+`composition' is included in the returned list.\n\
+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);
-
- if (! STRING_MULTIBYTE (str))
- return Qnil;
+ multibyte = STRING_MULTIBYTE (str);
bzero (charsets, (MAX_CHARSET + 1) * sizeof (int));
find_charset_in_str (XSTRING (str)->data, STRING_BYTES (XSTRING (str)),
- charsets, table, 0);
+ charsets, table, 1, 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) || i == CHARSET_COMPOSITION)
+ 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: %d %d", c1, c2);
+ c1 &= 0x7F;
+ c2 &= 0x7F;
+ if (c1 == 0
+ ? c2 != 0
+ : (c2 == 0
+ ? !CHAR_COMPONENTS_VALID_P (charset, c1, 0x20)
+ : !CHAR_COMPONENTS_VALID_P (charset, c1, c2)))
+ error ("Invalid code points: %d %d", 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),
}
DEFUN ("charset-after", Fcharset_after, Scharset_after, 0, 1, 0,
- "Return charset of a character in current buffer at position POS.\n\
-If POS is nil, it defauls to the current point.")
+ "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, c, charset;
+ 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);
+ {
+ 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);
- c = STRING_CHAR (p, Z_BYTE - pos_byte);
- charset = CHAR_CHARSET (c);
+ 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);
}
if (SINGLE_BYTE_CHAR_P (c))
return 1;
SPLIT_NON_ASCII_CHAR (c, charset, c1, c2);
- if (!CHARSET_VALID_P (charset))
- return 0;
- return (c < MIN_CHAR_COMPOSITION
- ? ((c & CHAR_FIELD1_MASK) /* i.e. dimension of C is two. */
- ? (genericp && c1 == 0 && c2 == 0
- || c1 >= 32 && c2 >= 32)
- : (genericp && c1 == 0
- || c1 >= 32))
- : c < MIN_CHAR_COMPOSITION + n_cmpchars);
+ if (charset == CHARSET_COMPOSITION)
+ return ((c >= MIN_CHAR_COMPOSITION
+ && c < MIN_CHAR_COMPOSITION + n_cmpchars)
+ || (genericp && c == GENERIC_COMPOSITION_CHAR));
+ 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 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)))
+ return make_number (1);
+}
+
+/* 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 bytes;
+
+ if (SINGLE_BYTE_CHAR_P (c) || (c & ~GLYPH_MASK_CHAR))
+ return 1;
+
+ if (COMPOSITE_CHAR_P (c))
{
- unsigned int id = COMPOSITE_CHAR_ID (XFASTINT (ch));
+ unsigned int id = COMPOSITE_CHAR_ID (c);
bytes = (id < n_cmpchars ? cmpchar_table[id]->len : 1);
}
else
{
- int charset = CHAR_CHARSET (XFASTINT (ch));
+ int charset = CHAR_CHARSET (c);
bytes = CHARSET_DEFINED_P (charset) ? CHARSET_BYTES (charset) : 1;
}
- XSETFASTINT (val, bytes);
- return val;
+ return bytes;
}
/* Return the width of character of which multi-byte form starts with
else if (COMPOSITE_CHAR_P (c))
{
int id = COMPOSITE_CHAR_ID (XFASTINT (ch));
- XSETFASTINT (val, (id < n_cmpchars ? cmpchar_table[id]->width : 0));
+ XSETFASTINT (val, (id < n_cmpchars ? cmpchar_table[id]->width : 1));
}
else
{
{
int from, to;
+ CHECK_NUMBER_COERCE_MARKER (beg, 0);
+ CHECK_NUMBER_COERCE_MARKER (end, 1);
+
from = min (XFASTINT (beg), XFASTINT (end));
to = max (XFASTINT (beg), XFASTINT (end));
unsigned char *ptr;
int nbytes;
{
- unsigned char *endp, c;
- int chars;
-
/* current_buffer is null at early stages of Emacs initialization. */
if (current_buffer == 0
|| NILP (current_buffer->enable_multibyte_characters))
return nbytes;
- endp = ptr + nbytes;
- chars = 0;
-
- while (ptr < endp)
- {
- c = *ptr++;
-
- if (BASE_LEADING_CODE_P (c))
- while (ptr < endp && ! CHAR_HEAD_P (*ptr)) ptr++;
- chars++;
- }
-
- return chars;
+ return multibyte_chars_in_text (ptr, nbytes);
}
/* Return the number of characters in the NBYTES bytes at PTR.
unsigned char *ptr;
int nbytes;
{
- unsigned char *endp, c;
- int chars;
+ unsigned char *endp;
+ int chars, bytes;
endp = ptr + nbytes;
chars = 0;
while (ptr < endp)
{
- c = *ptr++;
-
- if (BASE_LEADING_CODE_P (c))
- while (ptr < endp && ! CHAR_HEAD_P (*ptr)) ptr++;
+ if (BASE_LEADING_CODE_P (*ptr))
+ {
+ PARSE_MULTIBYTE_SEQ (ptr, nbytes, bytes);
+ ptr += bytes;
+ nbytes -= bytes;
+ }
+ else
+ {
+ ptr++;
+ nbytes--;
+ }
chars++;
}
p += len;
}
- val = make_string_from_bytes (buf, n, p - buf);
+ /* Here, we can't use make_string_from_bytes because of byte
+ combining problem. */
+ val = make_string (buf, p - buf);
return val;
}
int i;
struct cmpchar_info *cmpcharp;
- /* The second byte 0xFF means compostion rule is embedded. */
+ /* The second byte 0xFF means COMPOSITION rule is embedded. */
embedded_rule = (str[1] == 0xFF);
/* At first, get the actual length of the composite character. */
p = str + 1;
while (p < endp)
{
- if (embedded_rule) p++;
+ if (embedded_rule)
+ {
+ p++;
+ if (p >= endp)
+ return -1;
+ }
/* No need of checking if *P is 0xA0 because
BYTES_BY_CHAR_HEAD (0x80) surely returns 2. */
p += BYTES_BY_CHAR_HEAD (*p - 0x20);
}
/* We have to register the composite character in cmpchar_table. */
- if (n_cmpchars > (CHAR_FIELD2_MASK | CHAR_FIELD3_MASK))
+ if (n_cmpchars >= (CHAR_FIELD2_MASK | CHAR_FIELD3_MASK))
/* No, we have no more room for a new composite character. */
return -1;
/* 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), 0);
+ = 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);
return n_cmpchars++;
}
-/* Return the Nth element of the composite character C. */
+/* Return the Nth element of the composite character C. If NOERROR is
+ nonzero, return 0 on error condition (C is an invalid composite
+ charcter, or N is out of range). */
int
-cmpchar_component (c, n)
- unsigned int c, n;
+cmpchar_component (c, n, noerror)
+ int c, n, noerror;
{
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;
+ if (id < 0 || id >= n_cmpchars)
+ {
+ /* C is not a valid composite character. */
+ if (noerror) return 0;
+ error ("Invalid composite character: %d", c) ;
+ }
+ if (n >= cmpchar_table[id]->glyph_len)
+ {
+ /* No such component. */
+ if (noerror) return 0;
+ args_out_of_range (make_number (c), make_number (n));
+ }
/* No face data is stored in glyph code. */
return ((int) (cmpchar_table[id]->glyph[n]));
}
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;
+ "Return the Nth component character of composite character CHARACTER.")
+ (character, n)
+ Lisp_Object character, n;
{
- int c;
+ int id;
CHECK_NUMBER (character, 0);
- CHECK_NUMBER (idx, 1);
-
- if ((c = cmpchar_component (XINT (character), XINT (idx))) < 0)
- args_out_of_range (character, idx);
+ CHECK_NUMBER (n, 1);
- return make_number (c);
+ return (make_number (cmpchar_component (XINT (character), XINT (n), 0)));
}
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\
+ "Return the Nth composition rule of 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.")
+on the (N-1)th component.\n\
+If CHARACTER should be composed relatively or N is 0, return 255.")
(character, n)
Lisp_Object character, n;
{
- int id, i;
+ int id;
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)
+ if (XINT (n) < 0 || XINT (n) >= cmpchar_table[id]->glyph_len)
args_out_of_range (character, n);
- return make_number (cmpchar_table[id]->cmp_rule[i]);
+ return make_number (cmpchar_table[id]->cmp_rule
+ ? cmpchar_table[id]->cmp_rule[XINT (n)]
+ : 255);
}
DEFUN ("composite-char-composition-rule-p", Fcmpchar_cmp_rule_p,
i = 1;
while (p < pend)
{
- if (*p < 0x20 || *p == 127) /* control code */
+ if (*p < 0x20) /* control code */
error ("Invalid component character: %d", *p);
else if (*p < 0x80) /* ASCII */
{
LEADING_CODE_COMPOSITION, keep the remaining bytes
unchanged. */
p++;
+ if (*p == 255)
+ error ("Can't compose a rule-based composition character");
ptemp = p;
while (! CHAR_HEAD_P (*p)) p++;
+ if (str_cmpchar_id (ptemp - 1, p - ptemp + 1) < 0)
+ error ("Can't compose an invalid composition character");
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);
{
/* Add 0x20 to the base leading-code, keep the remaining
bytes unchanged. */
- len = BYTES_BY_CHAR_HEAD (*p);
+ int c = STRING_CHAR_AND_LENGTH (p, pend - p, len);
+
+ if (len <= 1 || ! CHAR_VALID_P (c, 0))
+ error ("Can't compose an invalid character");
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);
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);
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 (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 (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);
+ val = Fcons (make_number (GENERIC_COMPOSITION_CHAR), val);
+ Vgeneric_character_list = Fnreverse (val);
+ }
+
+ nonascii_insert_offset = 0;
+ Vnonascii_translation_table = Qnil;
}
#ifdef emacs
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 (&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);
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.
+If this is nil, `nonascii-insert-offset' is used instead.\n\
See also the docstring of `make-translation-table'.");
Vnonascii_translation_table = Qnil;
DEFVAR_INT ("min-composite-char", &min_composite_char,
"Minimum character code of a composite character.");
min_composite_char = MIN_CHAR_COMPOSITION;
+
+ 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 */