libregex: fix our mapping for \w

A small oversight in our implementation of \w is that it's actually
not strictly [[:alnum:]].  According to the GNU documentation, it's
actually [[:alnum:]] + underscore.  The fix is rather trivial: just add
it to our set explicitly, and amend our test set to be sure that _ is
actually included.

PR:		287396

lib/libc/regex/regcomp.c

@@ -1183,6 +1183,7 @@ p_b_pseudoclass(struct parse *p, char c) {
 		/* PASSTHROUGH */
 	case 'w':
 		p_b_cclass_named(p, cs, "alnum");
+		CHadd(p, cs, '_');
 		break;
 	case 'S':
 		cs->invert = 1;

lib/libregex/tests/gnuext.in

@@ -10,9 +10,9 @@ a\|b\|c	b	abc	a
 (ab)\1	-	abab	abab
 \1(ab)	C	ESUBREG
 (a)(b)(c)(d)(e)(f)(g)(h)(i)\9	-	abcdefghii	abcdefghii
-# \w, \W, \s, \S (alnum, ^alnum, space, ^space)
+# \w, \W, \s, \S (_alnum, ^_alnum, space, ^space)
-\w+	-	-%@a0X-	a0X
+\w+	-	-%@a_0X-	a_0X
-\w\+	b	-%@a0X-	a0X
+\w\+	b	-%@a_0X-	a_0X
 \s+	-	aSNTb	SNT
 \s\+	b	aSNTb	SNT
 # Word boundaries (\b, \B, \<, \>, \`, \')