tests: regcomp: assert character groups do match normal alphabet

In order to avoid us being unable to match characters which are part of
the normal US alphabet in certain weird languages, add two tests to
catch this behavior.

tests/core/posix.c

@@ -197,6 +197,48 @@ void test_core_posix__p_regcomp_ignores_global_locale_collate(void)
 	p_regfree(&preg);
 }
 
+void test_core_posix__p_regcomp_matches_digits_with_locale(void)
+{
+	p_regex_t preg;
+	char c, str[2];
+
+	try_set_locale(LC_COLLATE);
+	try_set_locale(LC_CTYPE);
+
+	cl_must_pass(p_regcomp(&preg, "[:digit:]", P_REG_EXTENDED));
+
+	str[1] = '\0';
+	for (c = '0'; c <= '9'; c++) {
+	    str[0] = c;
+	    cl_must_pass(p_regexec(&preg, str, 0, NULL, 0));
+	}
+
+	p_regfree(&preg);
+}
+
+void test_core_posix__p_regcomp_matches_alphabet_with_locale(void)
+{
+	p_regex_t preg;
+	char c, str[2];
+
+	try_set_locale(LC_COLLATE);
+	try_set_locale(LC_CTYPE);
+
+	cl_must_pass(p_regcomp(&preg, "[:alpha:]", REG_EXTENDED));
+
+	str[1] = '\0';
+	for (c = 'a'; c <= 'z'; c++) {
+	    str[0] = c;
+	    cl_must_pass(p_regexec(&preg, str, 0, NULL, 0));
+	}
+	for (c = 'A'; c <= 'Z'; c++) {
+	    str[0] = c;
+	    cl_must_pass(p_regexec(&preg, str, 0, NULL, 0));
+	}
+
+	p_regfree(&preg);
+}
+
 void test_core_posix__p_regcomp_compile_userdiff_regexps(void)
 {
 	size_t idx;