extcap_parser: remove G_REGEX_RAW from line parsing.

Check for utf8 valid line instead. Add a testcase that shows
how the former code was buggy on special characters extcap sentences.

Bug: 15668
Change-Id: Ic045c4791388af98705916e6ea84be8fc9b3c5b8
Reviewed-on: https://code.wireshark.org/review/32754
Petri-Dish: Dario Lombardo <lomato@gmail.com>
Tested-by: Petri Dish Buildbot
Reviewed-by: Peter Wu <peter@lekensteyn.nl>

extcap_parser.c

@@ -86,7 +86,9 @@ gdouble extcap_complex_get_double(extcap_complex *comp) {
 }
 
 static gboolean matches_regex(const char *pattern, const char *subject) {
-    return g_regex_match_simple(pattern, subject, (GRegexCompileFlags) (G_REGEX_CASELESS | G_REGEX_RAW), (GRegexMatchFlags)0);
+    if (!g_utf8_validate(subject, -1, NULL))
+        return FALSE;
+    return g_regex_match_simple(pattern, subject, (GRegexCompileFlags) (G_REGEX_CASELESS), (GRegexMatchFlags)0);
 }
 
 gboolean extcap_complex_get_bool(extcap_complex *comp) {
@@ -111,13 +113,16 @@ static extcap_token_sentence *extcap_tokenize_sentence(const gchar *s) {
     gchar *param_value = NULL;
     guint param_type = EXTCAP_PARAM_UNKNOWN;
 
+    if (!g_utf8_validate(s, -1, NULL))
+        return FALSE;
+
     extcap_token_sentence *rs = g_new0(extcap_token_sentence, 1);
 
     rs->sentence = NULL;
 
     /* Regex for catching just the allowed values for sentences */
     if ((regex = g_regex_new("^[\\t| ]*(arg|value|interface|extcap|dlt|control)(?=[\\t| ]+\\{)",
-                             (GRegexCompileFlags) (G_REGEX_CASELESS | G_REGEX_RAW),
+                             (GRegexCompileFlags) (G_REGEX_CASELESS),
                              (GRegexMatchFlags) 0, NULL)) != NULL) {
         g_regex_match(regex, s, (GRegexMatchFlags) 0, &match_info);
 
@@ -139,7 +144,7 @@ static extcap_token_sentence *extcap_tokenize_sentence(const gchar *s) {
      * that regex patterns given to {validation=} are parsed correctly,
      * as long as }{ does not occur within the pattern */
     regex = g_regex_new("\\{([a-zA-Z_-]*?)\\=(.*?)\\}(?=\\{|$|\\s)",
-                        (GRegexCompileFlags) (G_REGEX_CASELESS | G_REGEX_RAW),
+                        (GRegexCompileFlags) (G_REGEX_CASELESS),
                         (GRegexMatchFlags) 0, NULL);
     if (regex != NULL) {
         g_regex_match_full(regex, s, -1, 0, (GRegexMatchFlags) 0, &match_info, &error);

test/sampleif.py

@@ -0,0 +1,54 @@
+#!/usr/bin/env python3
+#
+# Wireshark test dummy extcap
+#
+# Copyright (c) 2018-2019 Peter Wu <peter@lekensteyn.nl>
+#
+# SPDX-License-Identifier: GPL-2.0-or-later
+#
+import argparse, os, sys
+
+parser = argparse.ArgumentParser()
+
+# Actions
+parser.add_argument('--extcap-interfaces', action='store_true')
+parser.add_argument('--extcap-dlts', action='store_true')
+parser.add_argument('--extcap-config', action='store_true')
+parser.add_argument('--capture', action='store_true')
+parser.add_argument('--extcap-version')
+
+parser.add_argument('--extcap-interface', metavar='IFACE')
+
+parser.add_argument('--extcap-capture-filter', metavar='CFILTER')
+parser.add_argument('--fifo', metavar='FIFO')
+
+
+def extcap_interfaces():
+    print("extcap {version=1.0}")
+    print("interface {value=sampleif}{display=Remote dumpcap}")
+
+
+def extcap_dlts():
+    # Required for the interface to show up in the interface list
+    print("dlt {number=147}{name=USER0}{display=Remote capture dependent DLT}")
+
+
+def extcap_config():
+    print("arg {number=0}{call=--test1}{display=Remote SSH server address}{type=string}{tooltip=bla}{required=true}{group=Server}")
+    print("arg {number=1}{call=--test2}{display=[7] Urządzenie kompozytowe USB}{type=string}{tooltip=X}{group=Capture}")
+
+
+def main():
+    args = parser.parse_args()
+    if args.extcap_interfaces:
+        return extcap_interfaces()
+
+    if args.extcap_dlts:
+        return extcap_dlts()
+    elif args.extcap_config:
+        return extcap_config()
+    else:
+        parser.error('Unsupported')
+        return 1
+
+sys.exit(main())

test/suite_clopts.py

@@ -10,10 +10,12 @@
 '''Command line option tests'''
 
 import json
+import sys
 import os.path
 import subprocess
 import subprocesstest
 import fixtures
+import shutil
 
 #glossaries = ('fields', 'protocols', 'values', 'decodes', 'defaultprefs', 'currentprefs')
 
@@ -279,3 +281,27 @@ class case_tshark_z_expert(subprocesstest.SubprocessTestCase):
         self.assertFalse(self.grepOutput('Errors'))
         self.assertFalse(self.grepOutput('Warns'))
         self.assertFalse(self.grepOutput('Chats'))
+
+
+@fixtures.mark_usefixtures('test_env')
+@fixtures.uses_fixtures
+class case_tshark_extcap(subprocesstest.SubprocessTestCase):
+    def test_tshark_extcap_interfaces(self, cmd_tshark, program_path):
+        # Script extcaps don't work with the current code on windows.
+        # https://www.wireshark.org/docs/wsdg_html_chunked/ChCaptureExtcap.html
+        # TODO: skip this test until it will get fixed.
+        if sys.platform == 'win32':
+            self.skipTest('FIXME extcap .py scripts needs special treatment on Windows')
+        extcap_tool = 'sampleif.py'
+        target_dir = os.path.join(program_path, 'extcap')
+        target_file = os.path.join(target_dir, extcap_tool)
+        source_file = os.path.join(os.path.dirname(__file__), extcap_tool)
+        os.makedirs(target_dir, exist_ok=True)
+        shutil.copy2(source_file, target_file)
+        # Ensure the test extcap_tool is properly loaded
+        self.assertRun((cmd_tshark, '-D'))
+        self.assertEqual(1, self.countOutput('sampleif'))
+        # Ensure tshark lists 2 interfaces in the preferences
+        self.assertRun((cmd_tshark, '-G', 'currentprefs'))
+        self.assertEqual(2, self.countOutput('extcap.sampleif.test'))
+        os.remove(target_file)