[Fix] Replace space like characters in headers with plain space

rules/regexp/headers.lua

@@ -862,14 +862,14 @@ reconf['SUBJECT_HAS_EXCLAIM'] = {
 }
 
 reconf['SUBJECT_ENDS_QUESTION'] = {
-  re = 'Subject=/\\?\\s*$/H',
+  re = 'Subject=/\\?\\s*$/Hu',
   description = 'Subject ends with a question',
   score = 1.0,
   group = 'headers'
 }
 
 reconf['SUBJECT_HAS_QUESTION'] = {
-  re = string.format('%s & !%s', 'Subject=/\\?/H', 'Subject=/\\?\\s*$/H'),
+  re = string.format('%s & !%s', 'Subject=/\\?/H', 'Subject=/\\?\\s*$/Hu'),
   description = 'Subject contains a question',
   score = 0.0,
   group = 'headers'

src/libmime/mime_headers.c

@@ -516,8 +516,14 @@ rspamd_mime_header_sanity_check (GString *str)
 
 	for (i = 0; i < str->len; i ++) {
 		t = str->str[i];
-		if (!((t & 0x80) || g_ascii_isgraph (t) || t == ' ')) {
-			str->str[i] = '?';
+		if (!((t & 0x80) || g_ascii_isgraph (t))) {
+			if (g_ascii_isspace (t)) {
+				/* Replace spaces characters with plain space */
+				str->str[i] = ' ';
+			}
+			else {
+				str->str[i] = '?';
+			}
 		}
 	}
 }

src/libserver/re_cache.c

@@ -809,9 +809,10 @@ rspamd_re_cache_exec_re (struct rspamd_task *task,
 
 			ret = rspamd_re_cache_process_regexp_data (rt, re,
 					task, scvec, lenvec, headerlist->len, raw);
-			msg_debug_re_task ("checking header %s regexp: %s -> %d",
+			msg_debug_re_task ("checking header %s regexp: %s=%*s -> %d",
 					re_class->type_data,
-					rspamd_regexp_get_pattern (re), ret);
+					rspamd_regexp_get_pattern (re),
+					(int)lenvec[0], scvec[0], ret);
 			g_free (scvec);
 			g_free (lenvec);
 		}