(regex_compile): Substitute whitespace_regexp for spaces, if it is nonzero.

(whitespace_regexp): New variable.
(re_set_whitespace_regexp): New function.

1 files changed, 71 insertions, 2 deletions

diff --git a/src/regex.c b/src/regex.c
index 36787238a32..459c0ab26b1 100644
--- a/src/regex.c
+++ b/src/regex.c
@@ -1250,7 +1250,7 @@ reg_syntax_t re_syntax_options;
 
 reg_syntax_t
 re_set_syntax (syntax)
-    reg_syntax_t syntax;
+     reg_syntax_t syntax;
 {
   reg_syntax_t ret = re_syntax_options;
 
@@ -1258,6 +1258,17 @@ re_set_syntax (syntax)
   return ret;
 }
 WEAK_ALIAS (__re_set_syntax, re_set_syntax)
+
+/* Regexp to use to replace spaces, or NULL meaning don't.  */
+static re_char *whitespace_regexp;
+
+void
+re_set_whitespace_regexp (regexp)
+     re_char *regexp;
+{
+  whitespace_regexp = regexp;
+}
+WEAK_ALIAS (__re_set_syntax, re_set_syntax)
 
 /* This table gives an error message for each of the error codes listed
    in regex.h.  Obviously the order here has to be same as there.
@@ -2436,6 +2447,15 @@ regex_compile (pattern, size, syntax, bufp)
   /* If the object matched can contain multibyte characters.  */
   const boolean multibyte = RE_MULTIBYTE_P (bufp);
 
+  /* Nonzero if we have pushed down into a subpattern.  */
+  int in_subpattern = 0;
+
+  /* These hold the values of p, pattern, and pend from the main
+     pattern when we have pushed into a subpattern.  */
+  re_char *main_p;
+  re_char *main_pattern;
+  re_char *main_pend;
+
 #ifdef DEBUG
   debug++;
   DEBUG_PRINT1 ("\nCompiling pattern: ");
@@ -2498,12 +2518,61 @@ regex_compile (pattern, size, syntax, bufp)
   begalt = b = bufp->buffer;
 
   /* Loop through the uncompiled pattern until we're at the end.  */
-  while (p != pend)
+  while (1)
     {
+      if (p == pend)
+        {
+          /* If this is the end of an included regexp,
+             pop back to the main regexp and try again.  */
+          if (in_subpattern)
+            {
+              in_subpattern = 0;
+              pattern = main_pattern;
+              p = main_p;
+              pend = main_pend;
+              continue;
+            }
+          /* If this is the end of the main regexp, we are done.  */
+          break;
+        }
+
       PATFETCH (c);
 
       switch (c)
         {
+        case ' ':
+          {
+            re_char *p1 = p;
+
+            /* If there's no special whitespace regexp, treat
+               spaces normally.  */
+            if (!whitespace_regexp)
+              goto normal_char;
+
+            /* Peek past following spaces.  */
+            while (p1 != pend)
+              {
+                if (*p1 != ' ')
+                  break;
+                p1++;
+              }
+            /* If the spaces are followed by a repetition op,
+               treat them normally.  */
+            if (p1 == pend
+                || (*p1 == '*' || *p1 == '+' || *p1 == '?'
+                    || (*p1 == '\\' && p1 + 1 != pend && p1[1] == '{')))
+              goto normal_char;
+
+            /* Replace the spaces with the whitespace regexp.  */
+            in_subpattern = 1;
+            main_p = p1;
+            main_pend = pend;
+            main_pattern = pattern;
+            p = pattern = whitespace_regexp;
+            pend = p + strlen (p);
+            break;
+          }    
+
         case '^':
           {
             if (   /* If at start of pattern, it's an operator.  */