dfilter: Improve grammar to parse functions

A function is grammatically an identifier that is followed by '(' and ')'
according to some rules. We should avoid assuming a token is a function
just because it matches a registered function name.

Before:
  Filter: foobar(http.user_agent) contains "UPDATE"
  dftest: Syntax error near "(".

After:
  Filter: foobar(http.user_agent) contains "UPDATE"
  dftest: The function 'foobar' does not exist.

This has the problem that a function cannot have the same name
as a protocol but that limitation already existed before.

epan/dfilter/dfilter-int.h

@@ -72,6 +72,9 @@ extern stnode_t *df_lval;
 void
 dfilter_fail(dfwork_t *dfw, const char *format, ...) G_GNUC_PRINTF(2, 3);
 
+void
+dfilter_parse_fail(dfwork_t *dfw, const char *format, ...) G_GNUC_PRINTF(2, 3);
+
 void
 add_deprecated_token(GPtrArray *deprecated, const char *token);
 
@@ -81,6 +84,9 @@ free_deprecated(GPtrArray *deprecated);
 void
 DfilterTrace(FILE *TraceFILE, char *zTracePrompt);
 
+stnode_t *
+dfilter_new_function(dfwork_t *dfw, const char *name);
+
 const char *tokenstr(int token);
 
 #endif

epan/dfilter/dfilter.c

@@ -37,18 +37,45 @@ static void*	ParserObj = NULL;
  */
 dfwork_t *global_dfw;
 
+static void
+dfilter_vfail(dfwork_t *dfw, const char *format, va_list args)
+{
+	/* If we've already reported one error, don't overwite it */
+	if (dfw->error_message != NULL)
+		return;
+
+	dfw->error_message = g_strdup_vprintf(format, args);
+}
+
 void
 dfilter_fail(dfwork_t *dfw, const char *format, ...)
 {
 	va_list	args;
 
-	/* If we've already reported one error, don't overwite it */
-	if (dfw->error_message != NULL)
-		return;
+	va_start(args, format);
+	dfilter_vfail(dfw, format, args);
+	va_end(args);
+}
+
+void
+dfilter_parse_fail(dfwork_t *dfw, const char *format, ...)
+{
+	va_list	args;
 
 	va_start(args, format);
-	dfw->error_message = g_strdup_vprintf(format, args);
+	dfilter_vfail(dfw, format, args);
 	va_end(args);
+	dfw->syntax_error = TRUE;
+}
+
+stnode_t *
+dfilter_new_function(dfwork_t *dfw, const char *name)
+{
+	df_func_def_t *def = df_func_lookup(name);
+	if (!def) {
+		dfilter_parse_fail(dfw, "Function '%s' does not exist", name);
+	}
+	return stnode_new(STTYPE_FUNCTION, def, name);
 }
 
 /* Initialize the dfilter module */
@@ -228,7 +255,6 @@ const char *tokenstr(int token)
 		case TOKEN_RBRACE:	return "RBRACE";
 		case TOKEN_WHITESPACE:	return "WHITESPACE";
 		case TOKEN_DOTDOT:	return "DOTDOT";
-		case TOKEN_FUNCTION:	return "FUNCTION";
 		case TOKEN_LPAREN:	return "LPAREN";
 		case TOKEN_RPAREN:	return "RPAREN";
 		default:		return "<unknown>";

epan/dfilter/dfunctions.c

@@ -318,7 +318,7 @@ df_functions[] = {
 
 /* Lookup a display filter function record by name */
 df_func_def_t*
-df_func_lookup(char *name)
+df_func_lookup(const char *name)
 {
     df_func_def_t *func_def;
 

epan/dfilter/dfunctions.h

@@ -37,6 +37,6 @@ typedef struct {
 } df_func_def_t;
 
 /* Return the function definition record for a function of named "name" */
-df_func_def_t* df_func_lookup(char *name);
+df_func_def_t* df_func_lookup(const char *name);
 
 #endif

epan/dfilter/grammar.lemon

@@ -104,14 +104,11 @@ any "error" symbols are shifted, if possible. */
 			hfinfo = (header_field_info *)stnode_data(TOKEN);
 			dfilter_fail(dfw, "Syntax error near \"%s\".", hfinfo->abbrev);
 			break;
-		case STTYPE_FUNCTION:
-			dfilter_fail(dfw, "The function \"%s\" was unexpected in this context.",
-				sttype_function_funcdef(TOKEN)->name);
-			break;
 		/* These aren't handed to use as terminal tokens from
 		   the scanner, so was can assert that we'll never
 		   see them here. */
 		case STTYPE_NUM_TYPES:
+		case STTYPE_FUNCTION:
 		case STTYPE_RANGE:
 		case STTYPE_FVALUE:
 		case STTYPE_PCRE:
@@ -331,16 +328,18 @@ set_node_list(L) ::= set_node_list(P) WHITESPACE entity(X) DOTDOT entity(Y).
 /* Functions */
 
 /* A function can have one or more parameters */
-function(G) ::= FUNCTION(F) LPAREN function_params(P) RPAREN.
+function(F) ::= UNPARSED(U) LPAREN function_params(P) RPAREN.
 {
-	G = F;
-	sttype_function_set_params(G, P);
+	F = dfilter_new_function(dfw, stnode_token_value(U));
+	stnode_free(U);
+	sttype_function_set_params(F, P);
 }
 
 /* A function can have zero parameters. */
-function(G) ::= FUNCTION(F) LPAREN RPAREN.
+function(F) ::= UNPARSED(U) LPAREN RPAREN.
 {
-	G = F;
+	F = dfilter_new_function(dfw, stnode_token_value(U));
+	stnode_free(U);
 }
 
 function_params(P) ::= entity(E).

epan/dfilter/scanner.l

@@ -86,7 +86,6 @@ DIAG_OFF_FLEX
 
 static int set_lval_str(int token, const char *token_value);
 static int set_lval_field(int token, header_field_info *hfinfo, const char *token_value);
-static int set_lval_func(int token, df_func_def_t *func, const char *token_value);
 static int set_lval_int(dfwork_t *dfw, int token, const char *token_value);
 static int simple(int token, const char *token_value);
 #define SIMPLE(token) simple(token, yytext)
@@ -401,7 +400,6 @@ static gboolean str_to_gint32(dfwork_t *dfw, const char *s, gint32* pint);
 [-+[:alnum:]_:]+([.][-+[:alnum:]_:]+)*[.]{0,2} {
 	/* Is it a field name or some other value (float, integer, bytes, ...)? */
 	header_field_info *hfinfo;
-	df_func_def_t *df_func_def;
 
 	/* Trailing dot is allowed for floats, but make sure that trailing ".."
 	 * is interpreted as a token on its own. */
@@ -422,11 +420,6 @@ static gboolean str_to_gint32(dfwork_t *dfw, const char *s, gint32* pint);
 		return set_lval_field(TOKEN_FIELD, hfinfo, yytext);
 	}
 
-	df_func_def = df_func_lookup(yytext);
-	if (df_func_def) {
-		/* Yes, it's a dfilter function */
-		return set_lval_func(TOKEN_FUNCTION, df_func_def, yytext);
-	}
 	/* No match, so treat it as an unparsed string */
 	return set_lval_str(TOKEN_UNPARSED, yytext);
 }
@@ -510,14 +503,6 @@ set_lval_field(int token, header_field_info *hfinfo, const char *token_value)
 	return token;
 }
 
-static int
-set_lval_func(int token, df_func_def_t *func, const char *token_value)
-{
-	ws_assert(token == TOKEN_FUNCTION);
-	stnode_init(df_lval, STTYPE_FUNCTION, func, token_value);
-	return token;
-}
-
 static int
 set_lval_int(dfwork_t *dfw, int token, const char *token_value)
 {

epan/dfilter/sttype-function.c

@@ -25,8 +25,6 @@ function_new(gpointer funcdef)
 {
 	function_t		*stfuncrec;
 
-	ws_assert(funcdef != NULL);
-
 	stfuncrec = g_new(function_t, 1);
 
 	stfuncrec->magic = FUNCTION_MAGIC;