dfilter: Fix parsing of value strings

If we have a STRING value in an expression and a numeric comparison
we must also check if it matches a value string before throwing
a type error.

Add appropriate tests to the test suite.

Fixes 4d2f469212.

epan/dfilter/semcheck.c

@@ -800,7 +800,7 @@ check_relation_LHS_FIELD(dfwork_t *dfw, const char *relation_string,
 			}
 
 			if (type2 == STTYPE_STRING) {
-				fvalue = dfilter_fvalue_from_string(dfw, ftype1, st_arg2, NULL);
+				fvalue = dfilter_fvalue_from_string(dfw, ftype1, st_arg2, hfinfo1);
 			}
 			else if (type2 == STTYPE_CHARCONST &&
 			    strcmp(relation_string, "contains") == 0) {

test/suite_dfilter/group_string_type.py

@@ -177,3 +177,11 @@ class case_string(unittest.TestCase):
     def test_contains_unicode(self, checkDFilterCount):
         dfilter = 'tcp.flags.str contains "·······AP···"'
         checkDFilterCount(dfilter, 1)
+
+    def test_value_string_1(self, checkDFilterCount):
+        dfilter = 'tcp.checksum.status == "Unverified" || tcp.checksum.status == "Good"'
+        checkDFilterCount(dfilter, 1)
+
+    def test_value_string_2(self, checkDFilterCount):
+        dfilter = 'tcp.checksum.status in {"Unverified" "Good"}'
+        checkDFilterCount(dfilter, 1)

test/suite_dfilter/group_syntax.py

@@ -26,3 +26,7 @@ class case_syntax(unittest.TestCase):
     def test_func_1(self, checkDFilterCount):
         dfilter = "len(frame) == 207"
         checkDFilterCount(dfilter, 1)
+
+    def test_value_string_1(self, checkDFilterSucceed):
+        dfilter = 'eth.fcs.status=="Bad"'
+        checkDFilterSucceed(dfilter)