/* GNU Emacs routines to deal with syntax tables; also word and list parsing.
- Copyright (C) 1985, 1987 Free Software Foundation, Inc.
+ Copyright (C) 1985, 1987, 1993 Free Software Foundation, Inc.
This file is part of GNU Emacs.
GNU Emacs is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
-the Free Software Foundation; either version 1, or (at your option)
+the Free Software Foundation; either version 2, or (at your option)
any later version.
GNU Emacs is distributed in the hope that it will be useful,
the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */
-#include "config.h"
+#include <config.h>
#include <ctype.h>
#include "lisp.h"
#include "commands.h"
Lisp_Object Qsyntax_table_p;
+static void scan_sexps_forward ();
+static int char_quoted ();
+
int words_include_escapes;
+/* This is the internal form of the parse state used in parse-partial-sexp. */
+
+struct lisp_parse_state
+ {
+ int depth; /* Depth at end of parsing */
+ int instring; /* -1 if not within string, else desired terminator. */
+ int incomment; /* Nonzero if within a comment at end of parsing */
+ int comstyle; /* comment style a=0, or b=1 */
+ int quoted; /* Nonzero if just after an escape char at end of parsing */
+ int thislevelstart; /* Char number of most recent start-of-expression at current level */
+ int prevlevelstart; /* Char number of start of containing expression */
+ int location; /* Char number at which parsing stopped. */
+ int mindepth; /* Minimum depth seen while scanning. */
+ int comstart; /* Position just after last comment starter. */
+ };
+\f
+/* These variables are a cache for finding the start of a defun.
+ find_start_pos is the place for which the defun start was found.
+ find_start_value is the defun start position found for it.
+ find_start_buffer is the buffer it was found in.
+ find_start_begv is the BEGV value when it was found.
+ find_start_modiff is the value of MODIFF when it was found. */
+
+static int find_start_pos;
+static int find_start_value;
+static struct buffer *find_start_buffer;
+static int find_start_begv;
+static int find_start_modiff;
+
+/* Find a defun-start that is the last one before POS (or nearly the last).
+ We record what we find, so that another call in the same area
+ can return the same value right away. */
+
+static int
+find_defun_start (pos)
+ int pos;
+{
+ int tem;
+ int shortage;
+
+ /* Use previous finding, if it's valid and applies to this inquiry. */
+ if (current_buffer == find_start_buffer
+ /* Reuse the defun-start even if POS is a little farther on.
+ POS might be in the next defun, but that's ok.
+ Our value may not be the best possible, but will still be usable. */
+ && pos <= find_start_pos + 1000
+ && pos >= find_start_value
+ && BEGV == find_start_begv
+ && MODIFF == find_start_modiff)
+ return find_start_value;
+
+ /* Back up to start of line. */
+ tem = scan_buffer ('\n', pos, -1, &shortage);
+
+ while (tem > BEGV)
+ {
+ /* Open-paren at start of line means we found our defun-start. */
+ if (SYNTAX (FETCH_CHAR (tem)) == Sopen)
+ break;
+ /* Move to beg of previous line. */
+ tem = scan_buffer ('\n', tem, -2, &shortage);
+ }
+
+ /* Record what we found, for the next try. */
+ find_start_value = tem;
+ find_start_buffer = current_buffer;
+ find_start_modiff = MODIFF;
+ find_start_begv = BEGV;
+ find_start_pos = pos;
+
+ return find_start_value;
+}
+\f
DEFUN ("syntax-table-p", Fsyntax_table_p, Ssyntax_table_p, 1, 1, 0,
"Return t if ARG is a syntax table.\n\
Any vector of 256 elements will do.")
register Lisp_Object tem;
while (tem = Fsyntax_table_p (obj),
NILP (tem))
- obj = wrong_type_argument (Qsyntax_table_p, obj, 0);
+ obj = wrong_type_argument (Qsyntax_table_p, obj);
return obj;
}
table = check_syntax_table (table);
current_buffer->syntax_table = table;
/* Indicate that this buffer now has a specified syntax table. */
- current_buffer->local_var_flags |= buffer_local_flags.syntax_table;
+ current_buffer->local_var_flags
+ |= XFASTINT (buffer_local_flags.syntax_table);
return table;
}
\f
0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
0377, 0377, 0377, 0377,
(char) Scomment, 0377, (char) Sendcomment, 0377,
- 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A, ... */
+ (char) Sinherit, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A ... */
0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword,
0377, 0377, 0377, 0377, (char) Sescape, 0377, 0377, (char) Ssymbol,
/* Indexed by syntax code, give the letter that describes it. */
-char syntax_code_spec[13] =
+char syntax_code_spec[14] =
{
- ' ', '.', 'w', '_', '(', ')', '\'', '\"', '$', '\\', '/', '<', '>'
+ ' ', '.', 'w', '_', '(', ')', '\'', '\"', '$', '\\', '/', '<', '>', '@'
};
\f
DEFUN ("char-syntax", Fchar_syntax, Schar_syntax, 1, 1, 0,
for make-docfile to see. We cannot put this in the real DEFUN
due to limits in the Unix cpp.
-DEFUN ("modify-syntax-entry", foo, bar, 0, 0, 0,
+DEFUN ("modify-syntax-entry", foo, bar, 2, 3, 0,
"Set syntax for character CHAR according to string S.\n\
The syntax is changed only for table TABLE, which defaults to\n\
the current buffer's syntax table.\n\
The first character of S should be one of the following:\n\
- Space whitespace syntax. w word constituent.\n\
- _ symbol constituent. . punctuation.\n\
- ( open-parenthesis. ) close-parenthesis.\n\
- \" string quote. \\ character-quote.\n\
- $ paired delimiter. ' expression quote or prefix operator.\n\
- < comment starter. > comment ender.\n\
+ Space or - whitespace syntax. w word constituent.\n\
+ _ symbol constituent. . punctuation.\n\
+ ( open-parenthesis. ) close-parenthesis.\n\
+ \" string quote. \\ escape.\n\
+ $ paired delimiter. ' expression quote or prefix operator.\n\
+ < comment starter. > comment ender.\n\
+ / character-quote. @ inherit from `standard-syntax-table'.\n\
+\n\
Only single-character comment start and end sequences are represented thus.\n\
Two-character sequences are represented as described below.\n\
The second character of S is the matching parenthesis,\n\
used only if the first character is `(' or `)'.\n\
Any additional characters are flags.\n\
-Defined flags are the characters 1, 2, 3, 4, and p.\n\
+Defined flags are the characters 1, 2, 3, 4, b, and p.\n\
1 means C is the start of a two-char comment start sequence.\n\
2 means C is the second character of such a sequence.\n\
3 means C is the start of a two-char comment end sequence.\n\
4 means C is the second character of such a sequence.\n\
- p means C is a prefix character for `backward-prefix-chars';
- such characters are treated as whitespace when they occur
+\n\
+There can be up to two orthogonal comment sequences. This is to support\n\
+language modes such as C++. By default, all comment sequences are of style\n\
+a, but you can set the comment sequence style to b (on the second character\n\
+of a comment-start, or the first character of a comment-end sequence) using\n\
+this flag:\n\
+ b means C is part of comment sequence b.\n\
+\n\
+ p means C is a prefix character for `backward-prefix-chars';\n\
+ such characters are treated as whitespace when they occur\n\
between expressions.")
-
+ (char, s, table)
*/
DEFUN ("modify-syntax-entry", Fmodify_syntax_entry, Smodify_syntax_entry, 2, 3,
case 'p':
XFASTINT (val) |= 1 << 20;
break;
+
+ case 'b':
+ XFASTINT (val) |= 1 << 21;
+ break;
}
XVECTOR (syntax_table)->contents[0xFF & XINT (c)] = val;
\f
/* Dump syntax table to buffer in human-readable format */
+static void
describe_syntax (value)
Lisp_Object value;
{
register enum syntaxcode code;
- char desc, match, start1, start2, end1, end2, prefix;
+ char desc, match, start1, start2, end1, end2, prefix, comstyle;
char str[2];
Findent_to (make_number (16), make_number (1));
end1 = (XINT (value) >> 18) & 1;
end2 = (XINT (value) >> 19) & 1;
prefix = (XINT (value) >> 20) & 1;
+ comstyle = (XINT (value) >> 21) & 1;
if ((int) code < 0 || (int) code >= (int) Smax)
{
if (prefix)
insert ("p", 1);
+ if (comstyle)
+ insert ("b", 1);
insert_string ("\twhich means: ");
insert_string ("comment"); break;
case Sendcomment:
insert_string ("endcomment"); break;
+ case Sinherit:
+ insert_string ("inherit"); break;
default:
insert_string ("invalid");
return;
if (match)
{
insert_string (", matches ");
-
- str[0] = match, str[1] = 0;
- insert (str, 1);
+ insert_char (match);
}
if (start1)
insert_string (",\n\t is the first character of a comment-end sequence");
if (end2)
insert_string (",\n\t is the second character of a comment-end sequence");
+ if (comstyle)
+ insert_string (" (comment style b)");
+
if (prefix)
insert_string (",\n\t is a prefix character for `backward-prefix-chars'");
insert_string ("\n");
}
-Lisp_Object
+static Lisp_Object
describe_syntax_1 (vector)
Lisp_Object vector;
{
struct buffer *old = current_buffer;
set_buffer_internal (XBUFFER (Vstandard_output));
- describe_vector (vector, Qnil, describe_syntax, 0, Qnil, Qnil);
+ describe_vector (vector, Qnil, describe_syntax, 0, Qnil);
set_buffer_internal (old);
return Qnil;
}
return Qt;
}
\f
+DEFUN ("forward-comment", Fforward_comment, Sforward_comment, 1, 1, 0,
+ "Move forward across up to N comments. If N is negative, move backward.\n\
+Stop scanning if we find something other than a comment or whitespace.\n\
+Set point to where scanning stops.\n\
+If N comments are found as expected, with nothing except whitespace\n\
+between them, return t; otherwise return nil.")
+ (count)
+ Lisp_Object count;
+{
+ register int from;
+ register int stop;
+ register int c;
+ register enum syntaxcode code;
+ int comstyle = 0; /* style of comment encountered */
+ int found;
+ int count1;
+
+ CHECK_NUMBER (count, 0);
+ count1 = XINT (count);
+
+ immediate_quit = 1;
+ QUIT;
+
+ from = PT;
+
+ while (count1 > 0)
+ {
+ stop = ZV;
+ while (from < stop)
+ {
+ c = FETCH_CHAR (from);
+ code = SYNTAX (c);
+ from++;
+ comstyle = 0;
+ if (from < stop && SYNTAX_COMSTART_FIRST (c)
+ && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from)))
+ {
+ /* We have encountered a comment start sequence and we
+ are ignoring all text inside comments. We must record
+ the comment style this sequence begins so that later,
+ only a comment end of the same style actually ends
+ the comment section. */
+ code = Scomment;
+ comstyle = SYNTAX_COMMENT_STYLE (FETCH_CHAR (from));
+ from++;
+ }
+
+ if (code == Scomment)
+ {
+ while (1)
+ {
+ if (from == stop)
+ {
+ immediate_quit = 0;
+ SET_PT (from);
+ return Qnil;
+ }
+ c = FETCH_CHAR (from);
+ if (SYNTAX (c) == Sendcomment
+ && SYNTAX_COMMENT_STYLE (c) == comstyle)
+ /* we have encountered a comment end of the same style
+ as the comment sequence which began this comment
+ section */
+ break;
+ from++;
+ if (from < stop && SYNTAX_COMEND_FIRST (c)
+ && SYNTAX_COMEND_SECOND (FETCH_CHAR (from))
+ && SYNTAX_COMMENT_STYLE (c) == comstyle)
+ /* we have encountered a comment end of the same style
+ as the comment sequence which began this comment
+ section */
+ { from++; break; }
+ }
+ /* We have skipped one comment. */
+ break;
+ }
+ else if (code != Swhitespace && code != Sendcomment)
+ {
+ immediate_quit = 0;
+ SET_PT (from - 1);
+ return Qnil;
+ }
+ }
+
+ /* End of comment reached */
+ count1--;
+ }
+
+ while (count1 < 0)
+ {
+ stop = BEGV;
+ while (from > stop)
+ {
+ int quoted;
+
+ from--;
+ quoted = char_quoted (from);
+ if (quoted)
+ from--;
+ c = FETCH_CHAR (from);
+ code = SYNTAX (c);
+ comstyle = 0;
+ if (code == Sendcomment)
+ comstyle = SYNTAX_COMMENT_STYLE (c);
+ if (from > stop && SYNTAX_COMEND_SECOND (c)
+ && SYNTAX_COMEND_FIRST (FETCH_CHAR (from - 1))
+ && !char_quoted (from - 1))
+ {
+ /* We must record the comment style encountered so that
+ later, we can match only the proper comment begin
+ sequence of the same style. */
+ code = Sendcomment;
+ comstyle = SYNTAX_COMMENT_STYLE (FETCH_CHAR (from - 1));
+ from--;
+ }
+
+ if (code == Sendcomment && !quoted)
+ {
+#if 0
+ if (code != SYNTAX (c))
+ /* For a two-char comment ender, we can assume
+ it does end a comment. So scan back in a simple way. */
+ {
+ if (from != stop) from--;
+ while (1)
+ {
+ if (SYNTAX (c = FETCH_CHAR (from)) == Scomment
+ && SYNTAX_COMMENT_STYLE (c) == comstyle)
+ break;
+ if (from == stop)
+ {
+ immediate_quit = 0;
+ SET_PT (from);
+ return Qnil;
+ }
+ from--;
+ if (SYNTAX_COMSTART_SECOND (c)
+ && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from))
+ && SYNTAX_COMMENT_STYLE (c) == comstyle
+ && !char_quoted (from))
+ break;
+ }
+ break;
+ }
+#endif /* 0 */
+
+ /* Look back, counting the parity of string-quotes,
+ and recording the comment-starters seen.
+ When we reach a safe place, assume that's not in a string;
+ then step the main scan to the earliest comment-starter seen
+ an even number of string quotes away from the safe place.
+
+ OFROM[I] is position of the earliest comment-starter seen
+ which is I+2X quotes from the comment-end.
+ PARITY is current parity of quotes from the comment end. */
+ {
+ int parity = 0;
+ char my_stringend = 0;
+ int string_lossage = 0;
+ int comment_end = from;
+ int comstart_pos = 0;
+ int comstart_parity = 0;
+
+ /* At beginning of range to scan, we're outside of strings;
+ that determines quote parity to the comment-end. */
+ while (from != stop)
+ {
+ /* Move back and examine a character. */
+ from--;
+
+ c = FETCH_CHAR (from);
+ code = SYNTAX (c);
+
+ /* If this char is the second of a 2-char comment sequence,
+ back up and give the pair the appropriate syntax. */
+ if (from > stop && SYNTAX_COMEND_SECOND (c)
+ && SYNTAX_COMEND_FIRST (FETCH_CHAR (from - 1)))
+ {
+ code = Sendcomment;
+ from--;
+ }
+
+ else if (from > stop && SYNTAX_COMSTART_SECOND (c)
+ && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from - 1))
+ && comstyle == SYNTAX_COMMENT_STYLE (c))
+ {
+ code = Scomment;
+ from--;
+ }
+
+ /* Ignore escaped characters. */
+ if (char_quoted (from))
+ continue;
+
+ /* Track parity of quotes. */
+ if (code == Sstring)
+ {
+ parity ^= 1;
+ if (my_stringend == 0)
+ my_stringend = c;
+ /* If we have two kinds of string delimiters.
+ There's no way to grok this scanning backwards. */
+ else if (my_stringend != c)
+ string_lossage = 1;
+ }
+
+ /* Record comment-starters according to that
+ quote-parity to the comment-end. */
+ if (code == Scomment)
+ {
+ comstart_parity = parity;
+ comstart_pos = from;
+ }
+
+ /* If we find another earlier comment-ender,
+ any comment-starts earlier than that don't count
+ (because they go with the earlier comment-ender). */
+ if (code == Sendcomment
+ && SYNTAX_COMMENT_STYLE (FETCH_CHAR (from)) == comstyle)
+ break;
+
+ /* Assume a defun-start point is outside of strings. */
+ if (code == Sopen
+ && (from == stop || FETCH_CHAR (from - 1) == '\n'))
+ break;
+ }
+
+ if (comstart_pos == 0)
+ from = comment_end;
+ /* If the earliest comment starter
+ is followed by uniform paired string quotes or none,
+ we know it can't be inside a string
+ since if it were then the comment ender would be inside one.
+ So it does start a comment. Skip back to it. */
+ else if (comstart_parity == 0 && !string_lossage)
+ from = comstart_pos;
+ else
+ {
+ /* We had two kinds of string delimiters mixed up
+ together. Decode this going forwards.
+ Scan fwd from the previous comment ender
+ to the one in question; this records where we
+ last passed a comment starter. */
+ struct lisp_parse_state state;
+ scan_sexps_forward (&state, find_defun_start (comment_end),
+ comment_end - 1, -10000, 0, Qnil, 0);
+ if (state.incomment)
+ from = state.comstart;
+ else
+ /* We can't grok this as a comment; scan it normally. */
+ from = comment_end;
+ }
+ }
+ /* We have skipped one comment. */
+ break;
+ }
+ else if ((code != Swhitespace && code != Scomment) || quoted)
+ {
+ immediate_quit = 0;
+ SET_PT (from + 1);
+ return Qnil;
+ }
+ }
+
+ count1++;
+ }
+
+ SET_PT (from);
+ immediate_quit = 0;
+ return Qt;
+}
+\f
int parse_sexp_ignore_comments;
Lisp_Object
int mathexit = 0;
register enum syntaxcode code;
int min_depth = depth; /* Err out if depth gets less than this. */
+ int comstyle = 0; /* style of comment encountered */
if (depth > 0) min_depth = 0;
while (from < stop)
{
c = FETCH_CHAR (from);
- code = SYNTAX(c);
+ code = SYNTAX (c);
from++;
if (from < stop && SYNTAX_COMSTART_FIRST (c)
&& SYNTAX_COMSTART_SECOND (FETCH_CHAR (from))
&& parse_sexp_ignore_comments)
- code = Scomment, from++;
+ {
+ /* we have encountered a comment start sequence and we
+ are ignoring all text inside comments. we must record
+ the comment style this sequence begins so that later,
+ only a comment end of the same style actually ends
+ the comment section */
+ code = Scomment;
+ comstyle = SYNTAX_COMMENT_STYLE (FETCH_CHAR (from));
+ from++;
+ }
+
if (SYNTAX_PREFIX (c))
continue;
while (from < stop)
{
#ifdef SWITCH_ENUM_BUG
- switch ((int) SYNTAX(FETCH_CHAR (from)))
+ switch ((int) SYNTAX (FETCH_CHAR (from)))
#else
- switch (SYNTAX(FETCH_CHAR (from)))
+ switch (SYNTAX (FETCH_CHAR (from)))
#endif
{
case Scharquote:
while (1)
{
if (from == stop) goto done;
- if (SYNTAX (c = FETCH_CHAR (from)) == Sendcomment)
+ c = FETCH_CHAR (from);
+ if (SYNTAX (c) == Sendcomment
+ && SYNTAX_COMMENT_STYLE (c) == comstyle)
+ /* we have encountered a comment end of the same style
+ as the comment sequence which began this comment
+ section */
break;
from++;
if (from < stop && SYNTAX_COMEND_FIRST (c)
- && SYNTAX_COMEND_SECOND (FETCH_CHAR (from)))
+ && SYNTAX_COMEND_SECOND (FETCH_CHAR (from))
+ && SYNTAX_COMMENT_STYLE (c) == comstyle)
+ /* we have encountered a comment end of the same style
+ as the comment sequence which began this comment
+ section */
{ from++; break; }
}
break;
if (from >= stop) goto lose;
if (FETCH_CHAR (from) == stringterm) break;
#ifdef SWITCH_ENUM_BUG
- switch ((int) SYNTAX(FETCH_CHAR (from)))
+ switch ((int) SYNTAX (FETCH_CHAR (from)))
#else
- switch (SYNTAX(FETCH_CHAR (from)))
+ switch (SYNTAX (FETCH_CHAR (from)))
#endif
{
case Scharquote:
from--;
c = FETCH_CHAR (from);
code = SYNTAX (c);
+ comstyle = 0;
+ if (code == Sendcomment)
+ comstyle = SYNTAX_COMMENT_STYLE (c);
if (from > stop && SYNTAX_COMEND_SECOND (c)
&& SYNTAX_COMEND_FIRST (FETCH_CHAR (from - 1))
&& !char_quoted (from - 1)
&& parse_sexp_ignore_comments)
- code = Sendcomment, from--;
+ {
+ /* we must record the comment style encountered so that
+ later, we can match only the proper comment begin
+ sequence of the same style */
+ code = Sendcomment;
+ comstyle = SYNTAX_COMMENT_STYLE (FETCH_CHAR (from - 1));
+ from--;
+ }
+
if (SYNTAX_PREFIX (c))
continue;
quoted = char_quoted (from - 1);
if (quoted)
from--;
- if (! (quoted || SYNTAX(FETCH_CHAR (from - 1)) == Sword
- || SYNTAX(FETCH_CHAR (from - 1)) == Ssymbol
- || SYNTAX(FETCH_CHAR (from - 1)) == Squote))
+ if (! (quoted || SYNTAX (FETCH_CHAR (from - 1)) == Sword
+ || SYNTAX (FETCH_CHAR (from - 1)) == Ssymbol
+ || SYNTAX (FETCH_CHAR (from - 1)) == Squote))
goto done2;
from--;
}
case Sendcomment:
if (!parse_sexp_ignore_comments)
break;
+#if 0
+ if (code != SYNTAX (c))
+ /* For a two-char comment ender, we can assume
+ it does end a comment. So scan back in a simple way. */
+ {
+ if (from != stop) from--;
+ while (1)
+ {
+ if (SYNTAX (c = FETCH_CHAR (from)) == Scomment
+ && SYNTAX_COMMENT_STYLE (c) == comstyle)
+ break;
+ if (from == stop) goto done;
+ from--;
+ if (SYNTAX_COMSTART_SECOND (c)
+ && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from))
+ && SYNTAX_COMMENT_STYLE (c) == comstyle
+ && !char_quoted (from))
+ break;
+ }
+ break;
+ }
+#endif /* 0 */
+
/* Look back, counting the parity of string-quotes,
and recording the comment-starters seen.
When we reach a safe place, assume that's not in a string;
which is I+2X quotes from the comment-end.
PARITY is current parity of quotes from the comment end. */
{
- int ofrom[2];
int parity = 0;
-
- ofrom[0] = ofrom[1] = from;
+ char my_stringend = 0;
+ int string_lossage = 0;
+ int comment_end = from;
+ int comstart_pos = 0;
+ int comstart_parity = 0;
/* At beginning of range to scan, we're outside of strings;
that determines quote parity to the comment-end. */
back up and give the pair the appropriate syntax. */
if (from > stop && SYNTAX_COMEND_SECOND (c)
&& SYNTAX_COMEND_FIRST (FETCH_CHAR (from - 1)))
- code = Sendcomment, from--;
+ {
+ code = Sendcomment;
+ from--;
+ }
+
else if (from > stop && SYNTAX_COMSTART_SECOND (c)
- && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from - 1)))
- code = Scomment, from--;
+ && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from - 1))
+ && comstyle == SYNTAX_COMMENT_STYLE (c))
+ {
+ code = Scomment;
+ from--;
+ }
/* Ignore escaped characters. */
if (char_quoted (from))
continue;
- /* Track parity of quotes between here and comment-end. */
+ /* Track parity of quotes. */
if (code == Sstring)
- parity ^= 1;
+ {
+ parity ^= 1;
+ if (my_stringend == 0)
+ my_stringend = c;
+ /* If we have two kinds of string delimiters.
+ There's no way to grok this scanning backwards. */
+ else if (my_stringend != c)
+ string_lossage = 1;
+ }
/* Record comment-starters according to that
quote-parity to the comment-end. */
if (code == Scomment)
- ofrom[parity] = from;
+ {
+ comstart_parity = parity;
+ comstart_pos = from;
+ }
+
+ /* If we find another earlier comment-ender,
+ any comment-starts earlier than that don't count
+ (because they go with the earlier comment-ender). */
+ if (code == Sendcomment
+ && SYNTAX_COMMENT_STYLE (FETCH_CHAR (from)) == comstyle)
+ break;
- /* If we come to another comment-end,
- assume it's not inside a string.
- That determines the quote parity to the comment-end. */
- if (code == Sendcomment)
+ /* Assume a defun-start point is outside of strings. */
+ if (code == Sopen
+ && (from == stop || FETCH_CHAR (from - 1) == '\n'))
break;
}
- from = ofrom[parity];
+
+ if (comstart_pos == 0)
+ from = comment_end;
+ /* If the earliest comment starter
+ is followed by uniform paired string quotes or none,
+ we know it can't be inside a string
+ since if it were then the comment ender would be inside one.
+ So it does start a comment. Skip back to it. */
+ else if (comstart_parity == 0 && !string_lossage)
+ from = comstart_pos;
+ else
+ {
+ /* We had two kinds of string delimiters mixed up
+ together. Decode this going forwards.
+ Scan fwd from the previous comment ender
+ to the one in question; this records where we
+ last passed a comment starter. */
+ struct lisp_parse_state state;
+ scan_sexps_forward (&state, find_defun_start (comment_end),
+ comment_end - 1, -10000, 0, Qnil, 0);
+ if (state.incomment)
+ from = state.comstart;
+ else
+ /* We can't grok this as a comment; scan it normally. */
+ from = comment_end;
+ }
}
break;
/* NOTREACHED */
}
+static int
char_quoted (pos)
register int pos;
{
return Qnil;
}
\f
-struct lisp_parse_state
- {
- int depth; /* Depth at end of parsing */
- int instring; /* -1 if not within string, else desired terminator. */
- int incomment; /* Nonzero if within a comment at end of parsing */
- int quoted; /* Nonzero if just after an escape char at end of parsing */
- int thislevelstart; /* Char number of most recent start-of-expression at current level */
- int prevlevelstart; /* Char number of start of containing expression */
- int location; /* Char number at which parsing stopped. */
- int mindepth; /* Minimum depth seen while scanning. */
- };
-
/* Parse forward from FROM to END,
- assuming that FROM is the start of a function,
- and return a description of the state of the parse at END. */
-
-struct lisp_parse_state val_scan_sexps_forward;
-
-struct lisp_parse_state *
-scan_sexps_forward (from, end, targetdepth, stopbefore, oldstate)
+ assuming that FROM has state OLDSTATE (nil means FROM is start of function),
+ and return a description of the state of the parse at END.
+ If STOPBEFORE is nonzero, stop at the start of an atom.
+ If COMMENTSTOP is nonzero, stop at the start of a comment. */
+
+static void
+scan_sexps_forward (stateptr, from, end, targetdepth,
+ stopbefore, oldstate, commentstop)
+ struct lisp_parse_state *stateptr;
register int from;
int end, targetdepth, stopbefore;
Lisp_Object oldstate;
+ int commentstop;
{
struct lisp_parse_state state;
depth = 0;
state.instring = -1;
state.incomment = 0;
+ state.comstyle = 0; /* comment style a by default */
}
else
{
oldstate = Fcdr (oldstate);
tem = Fcar (oldstate);
start_quoted = !NILP (tem);
+
+ /* if the eight element of the list is nil, we are in comment
+ style a. if it is non-nil, we are in comment style b */
+ oldstate = Fcdr (oldstate);
+ oldstate = Fcdr (oldstate);
+ tem = Fcar (oldstate);
+ state.comstyle = !NILP (tem);
}
state.quoted = 0;
mindepth = depth;
while (from < end)
{
- code = SYNTAX(FETCH_CHAR (from));
+ code = SYNTAX (FETCH_CHAR (from));
from++;
- if (from < end && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from - 1))
- && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from)))
- code = Scomment, from++;
+ if (code == Scomment)
+ state.comstart = from-1;
+
+ else if (from < end && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from - 1))
+ && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from)))
+ {
+ /* Record the comment style we have entered so that only
+ the comment-end sequence of the same style actually
+ terminates the comment section. */
+ code = Scomment;
+ state.comstyle = SYNTAX_COMMENT_STYLE (FETCH_CHAR (from));
+ state.comstart = from-1;
+ from++;
+ }
+
if (SYNTAX_PREFIX (FETCH_CHAR (from - 1)))
continue;
#ifdef SWITCH_ENUM_BUG
while (from < end)
{
#ifdef SWITCH_ENUM_BUG
- switch ((int) SYNTAX(FETCH_CHAR (from)))
+ switch ((int) SYNTAX (FETCH_CHAR (from)))
#else
- switch (SYNTAX(FETCH_CHAR (from)))
+ switch (SYNTAX (FETCH_CHAR (from)))
#endif
{
case Scharquote:
case Scomment:
state.incomment = 1;
startincomment:
+ if (commentstop)
+ goto done;
while (1)
{
if (from == end) goto done;
- if (SYNTAX (prev = FETCH_CHAR (from)) == Sendcomment)
+ prev = FETCH_CHAR (from);
+ if (SYNTAX (prev) == Sendcomment
+ && SYNTAX_COMMENT_STYLE (prev) == state.comstyle)
+ /* Only terminate the comment section if the endcomment
+ of the same style as the start sequence has been
+ encountered. */
break;
from++;
if (from < end && SYNTAX_COMEND_FIRST (prev)
- && SYNTAX_COMEND_SECOND (FETCH_CHAR (from)))
+ && SYNTAX_COMEND_SECOND (FETCH_CHAR (from))
+ && SYNTAX_COMMENT_STYLE (prev) == state.comstyle)
+ /* Only terminate the comment section if the end-comment
+ sequence of the same style as the start sequence has
+ been encountered. */
{ from++; break; }
}
state.incomment = 0;
+ state.comstyle = 0; /* reset the comment style */
break;
case Sopen:
if (from >= end) goto done;
if (FETCH_CHAR (from) == state.instring) break;
#ifdef SWITCH_ENUM_BUG
- switch ((int) SYNTAX(FETCH_CHAR (from)))
+ switch ((int) SYNTAX (FETCH_CHAR (from)))
#else
- switch (SYNTAX(FETCH_CHAR (from)))
+ switch (SYNTAX (FETCH_CHAR (from)))
#endif
{
case Scharquote:
state.location = from;
immediate_quit = 0;
- val_scan_sexps_forward = state;
- return &val_scan_sexps_forward;
+ *stateptr = state;
}
/* This comment supplies the doc string for parse-partial-sexp,
for make-docfile to see. We cannot put this in the real DEFUN
due to limits in the Unix cpp.
-DEFUN ("parse-partial-sexp", Ffoo, Sfoo, 0, 0, 0,
+DEFUN ("parse-partial-sexp", Ffoo, Sfoo, 2, 6, 0,
"Parse Lisp syntax starting at FROM until TO; return status of parse at TO.\n\
Parsing stops at TO or when certain criteria are met;\n\
point is set to where parsing stops.\n\
If fifth arg STATE is omitted or nil,\n\
parsing assumes that FROM is the beginning of a function.\n\
-Value is a list of seven elements describing final state of parsing:\n\
- 1. depth in parens.\n\
- 2. character address of start of innermost containing list; nil if none.\n\
- 3. character address of start of last complete sexp terminated.\n\
- 4. non-nil if inside a string.\n\
+Value is a list of eight elements describing final state of parsing:\n\
+ 0. depth in parens.\n\
+ 1. character address of start of innermost containing list; nil if none.\n\
+ 2. character address of start of last complete sexp terminated.\n\
+ 3. non-nil if inside a string.\n\
(it is the character that will terminate the string.)\n\
- 5. t if inside a comment.\n\
- 6. t if following a quote character.\n\
- 7. the minimum paren-depth encountered during this scan.\n\
+ 4. t if inside a comment.\n\
+ 5. t if following a quote character.\n\
+ 6. the minimum paren-depth encountered during this scan.\n\
+ 7. t if in a comment of style `b'.\n\
If third arg TARGETDEPTH is non-nil, parsing stops if the depth\n\
in parentheses becomes equal to TARGETDEPTH.\n\
Fourth arg STOPBEFORE non-nil means stop when come to\n\
any character that starts a sexp.\n\
-Fifth arg STATE is a seven-list like what this function returns.\n\
-It is used to initialize the state of the parse.")
-
+Fifth arg STATE is an eight-list like what this function returns.\n\
+It is used to initialize the state of the parse. Its second and third
+elements are ignored.
+Sixth args COMMENTSTOP non-nil means stop at the start of a comment.")
+ (from, to, targetdepth, stopbefore, state, commentstop)
*/
-DEFUN ("parse-partial-sexp", Fparse_partial_sexp, Sparse_partial_sexp, 2, 5, 0,
+DEFUN ("parse-partial-sexp", Fparse_partial_sexp, Sparse_partial_sexp, 2, 6, 0,
0 /* See immediately above */)
- (from, to, targetdepth, stopbefore, oldstate)
- Lisp_Object from, to, targetdepth, stopbefore, oldstate;
+ (from, to, targetdepth, stopbefore, oldstate, commentstop)
+ Lisp_Object from, to, targetdepth, stopbefore, oldstate, commentstop;
{
struct lisp_parse_state state;
int target;
target = -100000; /* We won't reach this depth */
validate_region (&from, &to);
- state = *scan_sexps_forward (XINT (from), XINT (to),
- target, !NILP (stopbefore), oldstate);
+ scan_sexps_forward (&state, XINT (from), XINT (to),
+ target, !NILP (stopbefore), oldstate,
+ !NILP (commentstop));
SET_PT (state.location);
Fcons (state.instring >= 0 ? make_number (state.instring) : Qnil,
Fcons (state.incomment ? Qt : Qnil,
Fcons (state.quoted ? Qt : Qnil,
- Fcons (make_number (state.mindepth), Qnil)))))));
+ Fcons (make_number (state.mindepth),
+ Fcons (state.comstyle ? Qt : Qnil,
+ Qnil))))))));
}
\f
init_syntax_once ()
defsubr (&Sforward_word);
+ defsubr (&Sforward_comment);
defsubr (&Sscan_lists);
defsubr (&Sscan_sexps);
defsubr (&Sbackward_prefix_chars);