/* Basic character support.
- Copyright (C) 1995, 1997, 1998, 2001 Electrotechnical Laboratory, JAPAN.
- Licensed to the Free Software Foundation.
- Copyright (C) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
- Free Software Foundation, Inc.
- Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
- National Institute of Advanced Industrial Science and Technology (AIST)
- Registration Number H13PRO009
+
+Copyright (C) 2001-2011 Free Software Foundation, Inc.
+Copyright (C) 1995, 1997, 1998, 2001 Electrotechnical Laboratory, JAPAN.
+ Licensed to the Free Software Foundation.
+Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
+ National Institute of Advanced Industrial Science and Technology (AIST)
+ Registration Number H13PRO009
This file is part of GNU Emacs.
Lisp_Object Qcharacterp;
-/* 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;
/* Char-table of information about which character to unify to which
Unicode character. Mainly used by the macro MAYBE_UNIFY_CHAR. */
Lisp_Object Vchar_unify_table;
-/* A char-table. An element is non-nil iff the corresponding
- character has a printable glyph. */
-Lisp_Object Vprintable_chars;
-
-/* A char-table. An elemnent is a column-width of the corresponding
- character. */
-Lisp_Object Vchar_width_table;
-
-/* A char-table. An element is a symbol indicating the direction
- property of corresponding character. */
-Lisp_Object Vchar_direction_table;
-
/* Variable used locally in the macro FETCH_MULTIBYTE_CHAR. */
unsigned char *_fetch_multibyte_char_p;
-/* Char table of scripts. */
-Lisp_Object Vchar_script_table;
-
-/* Alist of scripts vs representative characters. */
-Lisp_Object Vscript_representative_chars;
-
static Lisp_Object Qchar_script_table;
-Lisp_Object Vunicode_category_table;
\f
/* If character code C has modifier masks, reflect them to the
}
}
-DEFUN ("char-bytes", Fchar_bytes, Schar_bytes, 1, 1, 0,
- doc: /* Return 1 regardless of the argument CHAR.
-This is now an obsolete function. We keep it just for backward compatibility.
-usage: (char-bytes CHAR) */)
- (Lisp_Object ch)
-{
- CHECK_CHARACTER (ch);
- return make_number (1);
-}
-
DEFUN ("char-width", Fchar_width, Schar_width, 1, 1, 0,
doc: /* Return width of CHAR when displayed in the current buffer.
The width is measured by how many columns it occupies on the screen.
characters and bytes of the substring in *NCHARS and *NBYTES
respectively. */
-int
-c_string_width (const unsigned char *str, int len, int precision, int *nchars, int *nbytes)
+EMACS_INT
+c_string_width (const unsigned char *str, EMACS_INT len, int precision,
+ EMACS_INT *nchars, EMACS_INT *nbytes)
{
- int i = 0, i_byte = 0;
- int width = 0;
+ EMACS_INT i = 0, i_byte = 0;
+ EMACS_INT width = 0;
struct Lisp_Char_Table *dp = buffer_display_table ();
while (i_byte < len)
current buffer. The width is measured by how many columns it
occupies on the screen. */
-int
-strwidth (const unsigned char *str, int len)
+EMACS_INT
+strwidth (const char *str, EMACS_INT len)
{
- return c_string_width (str, len, -1, NULL, NULL);
+ return c_string_width ((const unsigned char *) str, len, -1, NULL, NULL);
}
/* Return width of Lisp string STRING when displayed in the current
PRECISION, and set number of characters and bytes of the substring
in *NCHARS and *NBYTES respectively. */
-int
-lisp_string_width (Lisp_Object string, int precision, int *nchars, int *nbytes)
+EMACS_INT
+lisp_string_width (Lisp_Object string, int precision,
+ EMACS_INT *nchars, EMACS_INT *nbytes)
{
- int len = SCHARS (string);
+ EMACS_INT len = SCHARS (string);
/* This set multibyte to 0 even if STRING is multibyte when it
contains only ascii and eight-bit-graphic, but that's
intentional. */
int multibyte = len < SBYTES (string);
unsigned char *str = SDATA (string);
- int i = 0, i_byte = 0;
- int width = 0;
+ EMACS_INT i = 0, i_byte = 0;
+ EMACS_INT width = 0;
struct Lisp_Char_Table *dp = buffer_display_table ();
while (i < len)
multibyte_chars_in_text (const unsigned char *ptr, EMACS_INT nbytes)
{
const unsigned char *endp = ptr + nbytes;
- int chars = 0;
+ EMACS_INT chars = 0;
while (ptr < endp)
{
- int len = MULTIBYTE_LENGTH (ptr, endp);
+ EMACS_INT len = MULTIBYTE_LENGTH (ptr, endp);
if (len == 0)
abort ();
represented by 2-byte in a multibyte text. */
void
-parse_str_as_multibyte (const unsigned char *str, int len, int *nchars, int *nbytes)
+parse_str_as_multibyte (const unsigned char *str, EMACS_INT len,
+ EMACS_INT *nchars, EMACS_INT *nbytes)
{
const unsigned char *endp = str + len;
- int n, chars = 0, bytes = 0;
+ EMACS_INT n, chars = 0, bytes = 0;
if (len >= MAX_MULTIBYTE_LENGTH)
{
area and that is enough. Return the number of bytes of the
resulting text. */
-int
-str_as_multibyte (unsigned char *str, int len, int nbytes, int *nchars)
+EMACS_INT
+str_as_multibyte (unsigned char *str, EMACS_INT len, EMACS_INT nbytes,
+ EMACS_INT *nchars)
{
unsigned char *p = str, *endp = str + nbytes;
unsigned char *to;
- int chars = 0;
+ EMACS_INT chars = 0;
int n;
if (nbytes >= MAX_MULTIBYTE_LENGTH)
bytes it may ocupy when converted to multibyte string by
`str_to_multibyte'. */
-int
-parse_str_to_multibyte (const unsigned char *str, int len)
+EMACS_INT
+parse_str_to_multibyte (const unsigned char *str, EMACS_INT len)
{
const unsigned char *endp = str + len;
- int bytes;
+ EMACS_INT bytes;
for (bytes = 0; str < endp; str++)
bytes += (*str < 0x80) ? 1 : 2;
that we can use LEN bytes at STR as a work area and that is
enough. */
-int
-str_to_multibyte (unsigned char *str, int len, int bytes)
+EMACS_INT
+str_to_multibyte (unsigned char *str, EMACS_INT len, EMACS_INT bytes)
{
unsigned char *p = str, *endp = str + bytes;
unsigned char *to;
actually converts characters in the range 0x80..0xFF to
unibyte. */
-int
-str_as_unibyte (unsigned char *str, int bytes)
+EMACS_INT
+str_as_unibyte (unsigned char *str, EMACS_INT bytes)
{
const unsigned char *p = str, *endp = str + bytes;
unsigned char *to;
}
-int
+EMACS_INT
string_count_byte8 (Lisp_Object string)
{
int multibyte = STRING_MULTIBYTE (string);
- int nbytes = SBYTES (string);
+ EMACS_INT nbytes = SBYTES (string);
unsigned char *p = SDATA (string);
unsigned char *pend = p + nbytes;
- int count = 0;
+ EMACS_INT count = 0;
int c, len;
if (multibyte)
Lisp_Object
string_escape_byte8 (Lisp_Object string)
{
- int nchars = SCHARS (string);
- int nbytes = SBYTES (string);
+ EMACS_INT nchars = SCHARS (string);
+ EMACS_INT nbytes = SBYTES (string);
int multibyte = STRING_MULTIBYTE (string);
- int byte8_count;
+ EMACS_INT byte8_count;
const unsigned char *src, *src_end;
unsigned char *dst;
Lisp_Object val;
return string;
if (multibyte)
- /* Convert 2-byte sequence of byte8 chars to 4-byte octal. */
- val = make_uninit_multibyte_string (nchars + byte8_count * 3,
- nbytes + byte8_count * 2);
+ {
+ if ((MOST_POSITIVE_FIXNUM - nchars) / 3 < byte8_count
+ || (MOST_POSITIVE_FIXNUM - nbytes) / 2 < byte8_count)
+ error ("Maximum string size exceeded");
+
+ /* Convert 2-byte sequence of byte8 chars to 4-byte octal. */
+ val = make_uninit_multibyte_string (nchars + byte8_count * 3,
+ nbytes + byte8_count * 2);
+ }
else
- /* Convert 1-byte sequence of byte8 chars to 4-byte octal. */
- val = make_uninit_string (nbytes + byte8_count * 3);
+ {
+ if ((MOST_POSITIVE_FIXNUM - nchars) / 3 < byte8_count)
+ error ("Maximum string size exceeded");
+ /* Convert 1-byte sequence of byte8 chars to 4-byte octal. */
+ val = make_uninit_string (nbytes + byte8_count * 3);
+ }
src = SDATA (string);
src_end = src + nbytes;
defsubr (&Scharacterp);
defsubr (&Sunibyte_char_to_multibyte);
defsubr (&Smultibyte_char_to_unibyte);
- defsubr (&Schar_bytes);
defsubr (&Schar_width);
defsubr (&Sstring_width);
defsubr (&Schar_direction);
defsubr (&Schar_resolve_modifiers);
defsubr (&Sget_byte);
- DEFVAR_LISP ("translation-table-vector", &Vtranslation_table_vector,
+ DEFVAR_LISP ("translation-table-vector", Vtranslation_table_vector,
doc: /*
Vector recording all translation tables ever defined.
Each element is a pair (SYMBOL . TABLE) relating the table to the
symbol naming it. The ID of a translation table is an index into this vector. */);
Vtranslation_table_vector = Fmake_vector (make_number (16), Qnil);
- DEFVAR_LISP ("auto-fill-chars", &Vauto_fill_chars,
+ DEFVAR_LISP ("auto-fill-chars", Vauto_fill_chars,
doc: /*
A char-table for characters which invoke auto-filling.
Such characters have value t in this table. */);
CHAR_TABLE_SET (Vauto_fill_chars, ' ', Qt);
CHAR_TABLE_SET (Vauto_fill_chars, '\n', Qt);
- DEFVAR_LISP ("char-width-table", &Vchar_width_table,
+ DEFVAR_LISP ("char-width-table", Vchar_width_table,
doc: /*
A char-table for width (columns) of each character. */);
Vchar_width_table = Fmake_char_table (Qnil, make_number (1));
char_table_set_range (Vchar_width_table, MAX_5_BYTE_CHAR + 1, MAX_CHAR,
make_number (4));
- DEFVAR_LISP ("char-direction-table", &Vchar_direction_table,
+ DEFVAR_LISP ("char-direction-table", Vchar_direction_table,
doc: /* A char-table for direction of each character. */);
Vchar_direction_table = Fmake_char_table (Qnil, make_number (1));
- DEFVAR_LISP ("printable-chars", &Vprintable_chars,
+ DEFVAR_LISP ("printable-chars", Vprintable_chars,
doc: /* A char-table for each printable character. */);
Vprintable_chars = Fmake_char_table (Qnil, Qnil);
Fset_char_table_range (Vprintable_chars,
Fcons (make_number (160),
make_number (MAX_5_BYTE_CHAR)), Qt);
- DEFVAR_LISP ("char-script-table", &Vchar_script_table,
+ DEFVAR_LISP ("char-script-table", Vchar_script_table,
doc: /* Char table of script symbols.
It has one extra slot whose value is a list of script symbols. */);
Fput (Qchar_script_table, Qchar_table_extra_slots, make_number (1));
Vchar_script_table = Fmake_char_table (Qchar_script_table, Qnil);
- DEFVAR_LISP ("script-representative-chars", &Vscript_representative_chars,
+ DEFVAR_LISP ("script-representative-chars", Vscript_representative_chars,
doc: /* Alist of scripts vs the representative characters.
Each element is a cons (SCRIPT . CHARS).
SCRIPT is a symbol representing a script or a subgroup of a script.
This variable is used to find a font for a specific script. */);
Vscript_representative_chars = Qnil;
- DEFVAR_LISP ("unicode-category-table", &Vunicode_category_table,
+ DEFVAR_LISP ("unicode-category-table", Vunicode_category_table,
doc: /* Char table of Unicode's "General Category".
All Unicode characters have one of the following values (symbol):
Lu, Ll, Lt, Lm, Lo, Mn, Mc, Me, Nd, Nl, No, Pc, Pd, Ps, Pe, Pi, Pf, Po,
}
#endif /* emacs */
-
-/* arch-tag: b6665960-3c3d-4184-85cd-af4318197999
- (do not change this comment) */
HCoop / bpt / emacs.git / blobdiff