/* Header for multibyte character handler.
Copyright (C) 1995, 1997, 1998 Electrotechnical Laboratory, JAPAN.
Licensed to the Free Software Foundation.
+ Copyright (C) 2001 Free Software Foundation, Inc.
This file is part of GNU Emacs.
A character set ("charset" hereafter) is a meaningful collection
(i.e. language, culture, functionality, etc) of characters. Emacs
handles multiple charsets at once. Each charset corresponds to one
- of ISO charsets. Emacs identifies a charset by a unique
+ of the ISO charsets. Emacs identifies a charset by a unique
identification number, whereas ISO identifies a charset by a triplet
of DIMENSION, CHARS and FINAL-CHAR. So, hereafter, just saying
"charset" means an identification number (integer value).
- The value range of charset is 0x00, 0x81..0xFE. There are four
+ The value range of charsets is 0x00, 0x81..0xFE. There are four
kinds of charset depending on DIMENSION (1 or 2) and CHARS (94 or
96). For instance, a charset of DIMENSION2_CHARS94 contains 94x94
characters.
Within Emacs Lisp, a charset is treated as a symbol which has a
property `charset'. The property value is a vector containing
- various information about the charset. For readability of C codes,
+ various information about the charset. For readability of C code,
we use the following convention for C variable names:
charset_symbol: Emacs Lisp symbol of a charset
charset_id: Emacs Lisp integer of an identification number of a charset
(range 0x80..0x9E). In addition, a charset of greater than 0xA0
(whose base leading-code is 0x9A..0x9D) is assigned an extended
leading-code (range 0xA0..0xFE). In this case, each base
- leading-code specify the allowable range of extended leading-code as
- shown in the table below. A leading-code is used to represent a
+ leading-code specifies the allowable range of extended leading-code
+ as shown in the table below. A leading-code is used to represent a
character in Emacs' buffer and string.
- We call a charset which has extended leading-code as "private
+ We call a charset which has extended leading-code a "private
charset" because those are mainly for a charset which is not yet
registered by ISO. On the contrary, we call a charset which does
- not have extended leading-code as "official charset".
+ not have extended leading-code an "official charset".
---------------------------------------------------------------------------
charset dimension base leading-code extended leading-code
/*** GENERAL NOTE on CHARACTER REPRESENTATION ***
- At first, the term "character" or "char" is used for a multilingual
- character (of course, including ASCII character), not for a byte in
+ Firstly, the term "character" or "char" is used for a multilingual
+ character (of course, including ASCII characters), not for a byte in
computer memory. We use the term "code" or "byte" for the latter
case.
POSITION-CODE is 0x20..0x7F.
Emacs has two kinds of representation of a character: multi-byte
- form (for buffer and string) and single-word form (for character
- object in Emacs Lisp). The latter is called "character code" here
- after. Both representations encode the information of charset and
- POSITION-CODE but in a different way (for instance, MSB of
+ form (for buffers and strings) and single-word form (for character
+ objects in Emacs Lisp). The latter is called "character code"
+ hereafter. Both representations encode the information of charset
+ and POSITION-CODE but in a different way (for instance, the MSB of
POSITION-CODE is set in multi-byte form).
- For details of multi-byte form, see the section "2. Emacs internal
- format handlers" of `coding.c'.
+ For details of the multi-byte form, see the section "2. Emacs
+ internal format handlers" of `coding.c'.
Emacs uses 19 bits for a character code. The bits are divided into
3 fields: FIELD1(5bits):FIELD2(7bits):FIELD3(7bits).
/* 1 if BYTE is an ASCII character in itself, in multibyte mode. */
#define ASCII_BYTE_P(byte) ((byte) < 0x80)
-/* A char-table containing information of each character set.
+/* A char-table containing information on each character set.
- Unlike ordinary char-tables, this doesn't contain any nested table.
+ Unlike ordinary char-tables, this doesn't contain any nested tables.
Only the top level elements are used. Each element is a vector of
the following information:
CHARSET-ID, BYTES, DIMENSION, CHARS, WIDTH, DIRECTION,
CHARSET-ID (integer) is the identification number of the charset.
- BYTES (integer) is the length of multi-byte form of a character in
- the charset: one of 1, 2, 3, and 4.
+ BYTES (integer) is the length of the multi-byte form of a character
+ in the charset: one of 1, 2, 3, and 4.
DIMENSION (integer) is the number of bytes to represent a character: 1 or 2.
charset.
LEADING-CODE-EXT (integer) is the extended leading-code for the
- charset. All charsets of less than 0xA0 has the value 0.
+ charset. All charsets of less than 0xA0 have the value 0.
ISO-FINAL-CHAR (character) is the final character of the
corresponding ISO 2022 charset. It is -1 for such a character
REVERSE-CHARSET (integer) is the charset which differs only in
LEFT-TO-RIGHT value from the charset. If there's no such a
charset, the value is -1.
-
+
SHORT-NAME (string) is the short name to refer to the charset.
LONG-NAME (string) is the long name to refer to the charset.
DESCRIPTION (string) is the description string of the charset.
PLIST (property list) may contain any type of information a user
- want to put and get by functions `put-charset-property' and
+ wants to put and get by functions `put-charset-property' and
`get-charset-property' respectively. */
extern Lisp_Object Vcharset_table;
#define UNIBYTE_STR_AS_MULTIBYTE_P(str, length, bytes) \
(((str)[0] < 0x80 || (str)[0] >= 0xA0) \
- ? (bytes) = 1 \
+ ? ((bytes) = 1) \
: (((bytes) = BYTES_BY_CHAR_HEAD ((str)[0])), \
((bytes) > 1 && (bytes) <= (length) \
&& (str)[0] != LEADING_CODE_8_BIT_CONTROL \
: char_bytes (c))
/* The following two macros CHAR_STRING and STRING_CHAR are the main
- entry points to convert between Emacs two types of character
+ entry points to convert between Emacs's two types of character
representations: multi-byte form and single-word form (character
code). */
: (*(str) = LEADING_CODE_8_BIT_CONTROL, *((str)+ 1) = c + 0x20, 2)) \
: char_to_string (c, (unsigned char *) str))
+/* Like CHAR_STRING but don't signal an error if C is invalid.
+ Value is -1 in this case. */
+
+#define CHAR_STRING_NO_SIGNAL(c, str) \
+ (SINGLE_BYTE_CHAR_P (c) \
+ ? ((ASCII_BYTE_P (c) || c >= 0xA0) \
+ ? (*(str) = (unsigned char)(c), 1) \
+ : (*(str) = LEADING_CODE_8_BIT_CONTROL, *((str)+ 1) = c + 0x20, 2)) \
+ : char_to_string_1 (c, (unsigned char *) str))
+
/* Return a character code of the character of which multi-byte form
is at STR and the length is LEN. If STR doesn't contain valid
multi-byte form, only the first byte in STR is returned. */
CHARIDX++; \
if (STRING_MULTIBYTE (STRING)) \
{ \
- unsigned char *ptr = &XSTRING (STRING)->data[BYTEIDX]; \
- int space_left = XSTRING (STRING)->size_byte - BYTEIDX; \
+ const unsigned char *ptr = SDATA (STRING) + BYTEIDX; \
+ int space_left = SBYTES (STRING) - BYTEIDX; \
int actual_len; \
\
OUTPUT = STRING_CHAR_AND_LENGTH (ptr, space_left, actual_len); \
BYTEIDX += actual_len; \
} \
else \
- OUTPUT = XSTRING (STRING)->data[BYTEIDX++]; \
+ OUTPUT = SREF (STRING, BYTEIDX++); \
} \
else
#define FETCH_STRING_CHAR_ADVANCE_NO_CHECK(OUTPUT, STRING, CHARIDX, BYTEIDX) \
if (1) \
{ \
- unsigned char *fetch_string_char_ptr = &XSTRING (STRING)->data[BYTEIDX]; \
- int fetch_string_char_space_left = XSTRING (STRING)->size_byte - BYTEIDX; \
+ const unsigned char *fetch_string_char_ptr = SDATA (STRING) + BYTEIDX; \
+ int fetch_string_char_space_left = SBYTES (STRING) - BYTEIDX; \
int actual_len; \
\
OUTPUT \
? 1 \
: multibyte_form_length (str, len))
+/* If P is before LIMIT, advance P to the next character boundary. It
+ assumes that P is already at a character boundary of the sane
+ mulitbyte form whose end address is LIMIT. */
+
+#define NEXT_CHAR_BOUNDARY(p, limit) \
+ do { \
+ if ((p) < (limit)) \
+ (p) += BYTES_BY_CHAR_HEAD (*(p)); \
+ } while (0)
+
+
+/* If P is after LIMIT, advance P to the previous character boundary.
+ It assumes that P is already at a character boundary of the sane
+ mulitbyte form whose beginning address is LIMIT. */
+
+#define PREV_CHAR_BOUNDARY(p, limit) \
+ do { \
+ if ((p) > (limit)) \
+ { \
+ const unsigned char *p0 = (p); \
+ do { \
+ p0--; \
+ } while (p0 >= limit && ! CHAR_HEAD_P (*p0)); \
+ (p) = (BYTES_BY_CHAR_HEAD (*p0) == (p) - p0) ? p0 : (p) - 1; \
+ } \
+ } while (0)
+
+
#ifdef emacs
/* Increase the buffer byte position POS_BYTE of the current buffer to
\
pos_byte--; \
if (pos_byte < GPT_BYTE) \
- p = BEG_ADDR + pos_byte - 1, p_min = BEG_ADDR; \
+ p = BEG_ADDR + pos_byte - BEG_BYTE, p_min = BEG_ADDR; \
else \
- p = BEG_ADDR + GAP_SIZE + pos_byte - 1, p_min = GAP_END_ADDR; \
+ p = BEG_ADDR + GAP_SIZE + pos_byte - BEG_BYTE, p_min = GAP_END_ADDR;\
if (p > p_min && !CHAR_HEAD_P (*p)) \
{ \
unsigned char *pend = p--; \
int len, bytes; \
+ if (p_min < p - MAX_MULTIBYTE_LENGTH) \
+ p_min = p - MAX_MULTIBYTE_LENGTH; \
while (p > p_min && !CHAR_HEAD_P (*p)) p--; \
len = pend + 1 - p; \
PARSE_MULTIBYTE_SEQ (p, len, bytes); \
pos_byte--; \
if (pos_byte < BUF_GPT_BYTE (buf)) \
{ \
- p = BUF_BEG_ADDR (buf) + pos_byte - 1; \
+ p = BUF_BEG_ADDR (buf) + pos_byte - BEG_BYTE; \
p_min = BUF_BEG_ADDR (buf); \
} \
else \
{ \
- p = BUF_BEG_ADDR (buf) + BUF_GAP_SIZE (buf) + pos_byte - 1; \
+ p = BUF_BEG_ADDR (buf) + BUF_GAP_SIZE (buf) + pos_byte - BEG_BYTE;\
p_min = BUF_GAP_END_ADDR (buf); \
} \
if (p > p_min && !CHAR_HEAD_P (*p)) \
{ \
unsigned char *pend = p--; \
int len, bytes; \
+ if (p_min < p - MAX_MULTIBYTE_LENGTH) \
+ p_min = p - MAX_MULTIBYTE_LENGTH; \
while (p > p_min && !CHAR_HEAD_P (*p)) p--; \
len = pend + 1 - p; \
PARSE_MULTIBYTE_SEQ (p, len, bytes); \
extern int split_string P_ ((const unsigned char *, int, int *,
unsigned char *, unsigned char *));
extern int char_to_string P_ ((int, unsigned char *));
+extern int char_to_string_1 P_ ((int, unsigned char *));
extern int string_to_char P_ ((const unsigned char *, int, int *));
extern int char_printable_p P_ ((int c));
extern int multibyte_form_length P_ ((const unsigned char *, int));
-extern void parse_str_as_multibyte P_ ((unsigned char *, int, int *, int *));
+extern void parse_str_as_multibyte P_ ((const unsigned char *, int, int *,
+ int *));
extern int str_as_multibyte P_ ((unsigned char *, int, int, int *));
+extern int parse_str_to_multibyte P_ ((unsigned char *, int));
extern int str_to_multibyte P_ ((unsigned char *, int, int));
extern int str_as_unibyte P_ ((unsigned char *, int));
extern int get_charset_id P_ ((Lisp_Object));
-extern int find_charset_in_text P_ ((unsigned char *, int, int, int *,
+extern int find_charset_in_text P_ ((const unsigned char *, int, int, int *,
Lisp_Object));
extern int strwidth P_ ((unsigned char *, int));
+extern int c_string_width P_ ((const unsigned char *, int, int, int *, int *));
+extern int lisp_string_width P_ ((Lisp_Object, int, int *, int *));
extern int char_bytes P_ ((int));
extern int char_valid_p P_ ((int, int));
+EXFUN (Funibyte_char_to_multibyte, 1);
+
extern Lisp_Object Vtranslation_table_vector;
/* Return a translation table of id number ID. */
#define BCOPY_SHORT(from, to, len) \
do { \
int i = len; \
- unsigned char *from_p = from, *to_p = to; \
+ const unsigned char *from_p = from; \
+ unsigned char *to_p = to; \
while (i--) *to_p++ = *from_p++; \
} while (0)