dfilter: require spaces as set element separator

Previously a filter such as `http.request.method in {"GET"HEAD""}` would
be parsed as three strings (GET, HEAD and an empty string). As it seems
more likely that people make typos rather than intending to construct
such a filter, forbid this by always requiring a whitespace separator.

Change-Id: I77e531fd6be072f62dd06aac27f856106c8920c6
Reported-by: Stig Bjørlykke
Reviewed-on: https://code.wireshark.org/review/26989
Petri-Dish: Peter Wu <peter@lekensteyn.nl>
Tested-by: Petri Dish Buildbot
Reviewed-by: Anders Broman <a.broman58@gmail.com>

4 files changed, 23 insertions, 6 deletions

diff --git a/epan/dfilter/dfilter-int.h b/epan/dfilter/dfilter-int.h
index 1f8dc0047a..785f5a3315 100644
--- a/epan/dfilter/dfilter-int.h
+++ b/epan/dfilter/dfilter-int.h
@@ -49,6 +49,7 @@ typedef struct {
 typedef struct {
 	dfwork_t *dfw;
 	GString* quoted_string;
+	gboolean in_set;	/* true if parsing set elements for the membership operator */
 } df_scanner_state_t;
 
 /* Constructor/Destructor prototypes for Lemon Parser */
diff --git a/epan/dfilter/dfilter.c b/epan/dfilter/dfilter.c
index a3ca2b6fcb..a975f84e84 100644
--- a/epan/dfilter/dfilter.c
+++ b/epan/dfilter/dfilter.c
@@ -239,6 +239,7 @@ dfilter_compile(const gchar *text, dfilter_t **dfp, gchar **err_msg)
 
 	state.dfw = dfw;
 	state.quoted_string = NULL;
+	state.in_set = FALSE;
 
 	df_set_extra(&state, scanner);
 
diff --git a/epan/dfilter/grammar.lemon b/epan/dfilter/grammar.lemon
index 647816e7a1..72ab5118ea 100644
--- a/epan/dfilter/grammar.lemon
+++ b/epan/dfilter/grammar.lemon
@@ -313,7 +313,7 @@ setnode_list(L) ::= entity(E).
 	L = g_slist_append(L, NULL);
 }
 
-setnode_list(L) ::= setnode_list(P) entity(E).
+setnode_list(L) ::= setnode_list(P) WHITESPACE entity(E).
 {
 	L = g_slist_append(P, E);
 	L = g_slist_append(L, NULL);
@@ -326,7 +326,7 @@ setnode_list(L) ::= entity(X) DOTDOT entity(Y).
 	L = g_slist_append(L, Y);
 }
 
-setnode_list(L) ::= setnode_list(P) entity(X) DOTDOT entity(Y).
+setnode_list(L) ::= setnode_list(P) WHITESPACE entity(X) DOTDOT entity(Y).
 {
 	L = g_slist_append(P, X);
 	L = g_slist_append(L, Y);
diff --git a/epan/dfilter/scanner.l b/epan/dfilter/scanner.l
index b53be6f1f3..735e37caef 100644
--- a/epan/dfilter/scanner.l
+++ b/epan/dfilter/scanner.l
@@ -111,16 +111,30 @@ static void mark_lval_deprecated(const char *s);
 
 %%
 
-[[:blank:]\n]+	/* ignore whitespace */
+[[:blank:]\n]+		{
+	/* Ignore whitespace, unless set elements are being parsed. Perhaps it
+	 * should have used commas from the beginning, but now we are stuck with
+	 * whitespace as separators. */
+	if (yyextra->in_set) {
+		return simple(TOKEN_WHITESPACE);
+	}
+}
 
 
 
 "("				return simple(TOKEN_LPAREN);
 ")"				return simple(TOKEN_RPAREN);
 ","				return simple(TOKEN_COMMA);
-"{"				return simple(TOKEN_LBRACE);
-".."				return simple(TOKEN_DOTDOT);
-"}"				return simple(TOKEN_RBRACE);
+
+"{"[[:blank:]\n]*	{
+	yyextra->in_set = TRUE;
+	return simple(TOKEN_LBRACE);
+}
+[[:blank:]\n]*".."[[:blank:]\n]*	return simple(TOKEN_DOTDOT);
+[[:blank:]\n]*"}"	{
+	yyextra->in_set = FALSE;
+	return simple(TOKEN_RBRACE);
+}
 
 "=="			return simple(TOKEN_TEST_EQ);
 "eq"			return simple(TOKEN_TEST_EQ);
@@ -397,6 +411,7 @@ simple(int token)
 		case TOKEN_COMMA:
 		case TOKEN_DOTDOT:
 		case TOKEN_HYPHEN:
+		case TOKEN_WHITESPACE:
 		case TOKEN_TEST_EQ:
 		case TOKEN_TEST_NE:
 		case TOKEN_TEST_GT: