1 files changed, 173 insertions, 23 deletions

diff --git a/src/syntax.c b/src/syntax.c
index 22a34c40fb3..57606f314d6 100644
--- a/src/syntax.c
+++ b/src/syntax.c
@@ -26,6 +26,7 @@ Boston, MA 02111-1307, USA.  */
 #include "buffer.h"
 #include "character.h"
 #include "keymap.h"
+#include "regex.h"
 
 /* Make syntax table lookup grant data in gl_state.  */
 #define SYNTAX_ENTRY_VIA_PROPERTY
@@ -97,12 +98,13 @@ static int find_start_modiff;
 static int find_defun_start P_ ((int, int));
 static int back_comment P_ ((int, int, int, int, int, int *, int *));
 static int char_quoted P_ ((int, int));
-static Lisp_Object skip_chars P_ ((int, Lisp_Object, Lisp_Object));
+static Lisp_Object skip_chars P_ ((int, Lisp_Object, Lisp_Object, int));
 static Lisp_Object skip_syntaxes P_ ((int, Lisp_Object, Lisp_Object));
 static Lisp_Object scan_lists P_ ((int, int, int, int));
 static void scan_sexps_forward P_ ((struct lisp_parse_state *,
                                     int, int, int, int,
                                     int, Lisp_Object, int));
+static int in_classes P_ ((int, Lisp_Object));
 
 
 struct gl_state_s gl_state;             /* Global state of syntax parser.  */
@@ -293,8 +295,11 @@ char_quoted (charpos, bytepos)
 
   while (bytepos >= beg)
     {
+      int c;
+
       UPDATE_SYNTAX_TABLE_BACKWARD (charpos);
-      code = SYNTAX (FETCH_CHAR_AS_MULTIBYTE (bytepos));
+      c = FETCH_CHAR_AS_MULTIBYTE (bytepos);
+      code = SYNTAX (c);
       if (! (code == Scharquote || code == Sescape))
         break;
 
@@ -381,12 +386,16 @@ find_defun_start (pos, pos_byte)
   gl_state.use_global = 0;
   while (PT > BEGV)
     {
+      int c;
+
       /* Open-paren at start of line means we may have found our
          defun-start.  */
-      if (SYNTAX (FETCH_CHAR_AS_MULTIBYTE (PT_BYTE)) == Sopen)
+      c = FETCH_CHAR_AS_MULTIBYTE (PT_BYTE);
+      if (SYNTAX (c) == Sopen)
         {
           SETUP_SYNTAX_TABLE (PT + 1, -1);      /* Try again... */
-          if (SYNTAX (FETCH_CHAR_AS_MULTIBYTE (PT_BYTE)) == Sopen)
+          c = FETCH_CHAR_AS_MULTIBYTE (PT_BYTE);
+          if (SYNTAX (c) == Sopen)
             break;
           /* Now fallback to the default value.  */
           gl_state.current_syntax_table = current_buffer->syntax_table;
@@ -955,7 +964,7 @@ text property.  */)
 DEFUN ("modify-syntax-entry", Fmodify_syntax_entry, Smodify_syntax_entry, 2, 3,
   "cSet syntax for character: \nsSet syntax for %s to: ",
        doc: /* Set syntax for character CHAR according to string NEWENTRY.
-The syntax is changed only for table SYNTAX_TABLE, which defaults to
+The syntax is changed only for table SYNTAX-TABLE, which defaults to
  the current buffer's syntax table.
 CHAR may be a cons (MIN . MAX), in which case, syntaxes of all characters
 in the range MIN and MAX are changed.
@@ -1339,13 +1348,13 @@ except that `]' is never special and `\\' quotes `^', `-' or `\\'
  (but not as the end of a range; quoting is never needed there).
 Thus, with arg "a-zA-Z", this skips letters stopping before first nonletter.
 With arg "^a-zA-Z", skips nonletters stopping before first letter.
-Returns the distance traveled, either zero or positive.
-Note that char classes, e.g. `[:alpha:]', are not currently supported;
-they will be treated as literals.  */)
+Char classes, e.g. `[:alpha:]', are supported.
+
+Returns the distance traveled, either zero or positive.  */)
      (string, lim)
      Lisp_Object string, lim;
 {
-  return skip_chars (1, string, lim);
+  return skip_chars (1, string, lim, 1);
 }
 
 DEFUN ("skip-chars-backward", Fskip_chars_backward, Sskip_chars_backward, 1, 2, 0,
@@ -1355,7 +1364,7 @@ Returns the distance traveled, either zero or negative.  */)
      (string, lim)
      Lisp_Object string, lim;
 {
-  return skip_chars (0, string, lim);
+  return skip_chars (0, string, lim, 1);
 }
 
 DEFUN ("skip-syntax-forward", Fskip_syntax_forward, Sskip_syntax_forward, 1, 2, 0,
@@ -1383,9 +1392,10 @@ This function returns the distance traveled, either zero or negative.  */)
 }
 
 static Lisp_Object
-skip_chars (forwardp, string, lim)
+skip_chars (forwardp, string, lim, handle_iso_classes)
      int forwardp;
      Lisp_Object string, lim;
+     int handle_iso_classes;
 {
   register unsigned int c;
   unsigned char fastmap[0400];
@@ -1403,8 +1413,10 @@ skip_chars (forwardp, string, lim)
   int size_byte;
   const unsigned char *str;
   int len;
+  Lisp_Object iso_classes;
 
   CHECK_STRING (string);
+  iso_classes = Qnil;
 
   if (NILP (lim))
     XSETINT (lim, forwardp ? ZV : BEGV);
@@ -1448,6 +1460,42 @@ skip_chars (forwardp, string, lim)
         {
           c = str[i_byte++];
 
+          if (handle_iso_classes && c == '['
+              && i_byte < size_byte
+              && str[i_byte] == ':')
+            {
+              const unsigned char *class_beg = str + i_byte + 1;
+              const unsigned char *class_end = class_beg;
+              const unsigned char *class_limit = str + size_byte - 2;
+              /* Leave room for the null.        */
+              unsigned char class_name[CHAR_CLASS_MAX_LENGTH + 1];
+              re_wctype_t cc;
+
+              if (class_limit - class_beg > CHAR_CLASS_MAX_LENGTH)
+                class_limit = class_beg + CHAR_CLASS_MAX_LENGTH;
+
+              while (class_end < class_limit
+                     && *class_end >= 'a' && *class_end <= 'z')
+                class_end++;
+
+              if (class_end == class_beg
+                  || *class_end != ':' || class_end[1] != ']')
+                goto not_a_class_name;
+
+              bcopy (class_beg, class_name, class_end - class_beg);
+              class_name[class_end - class_beg] = 0;
+
+              cc = re_wctype (class_name);
+              if (cc == 0)
+                error ("Invalid ISO C character class");
+
+              iso_classes = Fcons (make_number (cc), iso_classes);
+
+              i_byte = class_end + 2 - str;
+              continue;
+            }
+
+        not_a_class_name:
           if (c == '\\')
             {
               if (i_byte == size_byte)
@@ -1534,6 +1582,42 @@ skip_chars (forwardp, string, lim)
           c = STRING_CHAR_AND_LENGTH (str + i_byte, size_byte-i_byte, len);
           i_byte += len;
 
+          if (handle_iso_classes && c == '['
+              && i_byte < size_byte
+              && STRING_CHAR (str + i_byte, size_byte - i_byte) == ':')
+            {
+              const unsigned char *class_beg = str + i_byte + 1;
+              const unsigned char *class_end = class_beg;
+              const unsigned char *class_limit = str + size_byte - 2;
+              /* Leave room for the null.        */
+              unsigned char class_name[CHAR_CLASS_MAX_LENGTH + 1];
+              re_wctype_t cc;
+
+              if (class_limit - class_beg > CHAR_CLASS_MAX_LENGTH)
+                class_limit = class_beg + CHAR_CLASS_MAX_LENGTH;
+
+              while (class_end < class_limit
+                     && *class_end >= 'a' && *class_end <= 'z')
+                class_end++;
+
+              if (class_end == class_beg
+                  || *class_end != ':' || class_end[1] != ']')
+                goto not_a_class_name_multibyte;
+
+              bcopy (class_beg, class_name, class_end - class_beg);
+              class_name[class_end - class_beg] = 0;
+
+              cc = re_wctype (class_name);
+              if (cc == 0)
+                error ("Invalid ISO C character class");
+
+              iso_classes = Fcons (make_number (cc), iso_classes);
+
+              i_byte = class_end + 2 - str;
+              continue;
+            }
+
+        not_a_class_name_multibyte:
           if (c == '\\')
             {
               if (i_byte == size_byte)
@@ -1643,13 +1727,13 @@ skip_chars (forwardp, string, lim)
 
     if (forwardp)
       {
-        endp = (XINT (lim) == GPT) ? GPT_ADDR : CHAR_POS_ADDR (XINT (lim));
-        stop = (pos < GPT && GPT < XINT (lim)) ? GPT_ADDR : endp;
+        endp = (XINT (lim) == GPT) ? GPT_ADDR : CHAR_POS_ADDR (XINT (lim));
+        stop = (pos < GPT && GPT < XINT (lim)) ? GPT_ADDR : endp;
       }
     else
       {
-        endp = CHAR_POS_ADDR (XINT (lim));
-        stop = (pos >= GPT && GPT > XINT (lim)) ? GAP_END_ADDR : endp;
+        endp = CHAR_POS_ADDR (XINT (lim));
+        stop = (pos >= GPT && GPT > XINT (lim)) ? GAP_END_ADDR : endp;
       }
 
     immediate_quit = 1;
@@ -1667,9 +1751,17 @@ skip_chars (forwardp, string, lim)
                   p = GAP_END_ADDR;
                   stop = endp;
                 }
+              c = STRING_CHAR_AND_LENGTH (p, MAX_MULTIBYTE_LENGTH, nbytes);
+              if (! NILP (iso_classes) && in_classes (c, iso_classes))
+                {
+                  if (negate)
+                    break;
+                  else
+                    goto fwd_ok;
+                }
+
               if (! fastmap[*p])
                 break;
-              c = STRING_CHAR_AND_LENGTH (p, MAX_MULTIBYTE_LENGTH, nbytes);
               if (! ASCII_CHAR_P (c))
                 {
                   /* As we are looking at a multibyte character, we
@@ -1686,6 +1778,7 @@ skip_chars (forwardp, string, lim)
                   if (!(negate ^ (i < n_char_ranges)))
                     break;
                 }
+            fwd_ok:
               p += nbytes, pos++, pos_byte += nbytes;
             }
         else
@@ -1698,8 +1791,18 @@ skip_chars (forwardp, string, lim)
                   p = GAP_END_ADDR;
                   stop = endp;
                 }
+
+              if (!NILP (iso_classes) && in_classes (*p, iso_classes))
+                {
+                  if (negate)
+                    break;
+                  else
+                    goto fwd_unibyte_ok;
+                }
+
               if (!fastmap[*p])
                 break;
+            fwd_unibyte_ok:
               p++, pos++, pos_byte++;
             }
       }
@@ -1719,9 +1822,18 @@ skip_chars (forwardp, string, lim)
                 }
               prev_p = p;
               while (--p >= stop && ! CHAR_HEAD_P (*p));
+              c = STRING_CHAR (p, MAX_MULTIBYTE_LENGTH);
+
+              if (! NILP (iso_classes) && in_classes (c, iso_classes))
+                {
+                  if (negate)
+                    break;
+                  else
+                    goto back_ok;
+                }
+
               if (! fastmap[*p])
                 break;
-              c = STRING_CHAR (p, MAX_MULTIBYTE_LENGTH);
               if (! ASCII_CHAR_P (c))
                 {
                   /* See the comment in the previous similar code.  */
@@ -1731,6 +1843,7 @@ skip_chars (forwardp, string, lim)
                   if (!(negate ^ (i < n_char_ranges)))
                     break;
                 }
+            back_ok:
               pos--, pos_byte -= prev_p - p;
             }
         else
@@ -1743,8 +1856,18 @@ skip_chars (forwardp, string, lim)
                   p = GPT_ADDR;
                   stop = endp;
                 }
+
+              if (! NILP (iso_classes) && in_classes (p[-1], iso_classes))
+                {
+                  if (negate)
+                    break;
+                  else
+                    goto back_unibyte_ok;
+                }
+
               if (!fastmap[p[-1]])
                 break;
+            back_unibyte_ok:
               p--, pos--, pos_byte--;
             }
       }
@@ -1927,6 +2050,30 @@ skip_syntaxes (forwardp, string, lim)
     return make_number (PT - start_point);
   }
 }
+
+/* Return 1 if character C belongs to one of the ISO classes
+   in the list ISO_CLASSES.  Each class is represented by an
+   integer which is its type according to re_wctype.  */
+
+static int
+in_classes (c, iso_classes)
+     int c;
+     Lisp_Object iso_classes;
+{
+  int fits_class = 0;
+
+  while (! NILP (iso_classes))
+    {
+      Lisp_Object elt;
+      elt = XCAR (iso_classes);
+      iso_classes = XCDR (iso_classes);
+
+      if (re_iswctype (c, XFASTINT (elt)))
+        fits_class = 1;
+    }
+
+  return fits_class;
+}
 
 /* Jump over a comment, assuming we are at the beginning of one.
    FROM is the current position.
@@ -2310,7 +2457,8 @@ scan_lists (from, count, depth, sexpflag)
           INC_BOTH (from, from_byte);
           UPDATE_SYNTAX_TABLE_FORWARD (from);
           if (from < stop && comstart_first
-              && SYNTAX_COMSTART_SECOND (FETCH_CHAR_AS_MULTIBYTE (from_byte))
+              && (c = FETCH_CHAR_AS_MULTIBYTE (from_byte),
+                  SYNTAX_COMSTART_SECOND (c))
               && parse_sexp_ignore_comments)
             {
               /* we have encountered a comment start sequence and we
@@ -2636,7 +2784,7 @@ scan_lists (from, count, depth, sexpflag)
            Fcons (build_string ("Unbalanced parentheses"),
                   Fcons (make_number (last_good),
                          Fcons (make_number (from), Qnil))));
-
+  abort ();
   /* NOTREACHED */
 }
 
@@ -2776,8 +2924,8 @@ scan_sexps_forward (stateptr, from, from_byte, end, targetdepth,
 #define INC_FROM                                \
 do { prev_from = from;                          \
      prev_from_byte = from_byte;                \
-     prev_from_syntax                           \
-       = SYNTAX_WITH_FLAGS (FETCH_CHAR_AS_MULTIBYTE (prev_from_byte)); \
+     temp = FETCH_CHAR_AS_MULTIBYTE (prev_from_byte);   \
+     prev_from_syntax = SYNTAX_WITH_FLAGS (temp); \
      INC_BOTH (from, from_byte);                \
      if (from < end)                            \
        UPDATE_SYNTAX_TABLE_FORWARD (from);      \
@@ -2852,7 +3000,8 @@ do { prev_from = from;				\
   curlevel->last = -1;
 
   SETUP_SYNTAX_TABLE (prev_from, 1);
-  prev_from_syntax = SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte));
+  temp = FETCH_CHAR (prev_from_byte);
+  prev_from_syntax = SYNTAX_WITH_FLAGS (temp);
   UPDATE_SYNTAX_TABLE_FORWARD (from);
 
   /* Enter the loop at a place appropriate for initial state.  */
@@ -2931,7 +3080,8 @@ do { prev_from = from;				\
           while (from < end)
             {
               /* Some compilers can't handle this inside the switch.  */
-              temp = SYNTAX (FETCH_CHAR_AS_MULTIBYTE (from_byte));
+              temp = FETCH_CHAR_AS_MULTIBYTE (from_byte);
+              temp = SYNTAX (temp);
               switch (temp)
                 {
                 case Scharquote: