+m4_pushdef([AT_PREVIOUS_STATE_GRAMMAR],
+[[%nonassoc 'a';
+
+start: consistent-error-on-a-a 'a' ;
+
+consistent-error-on-a-a:
+ 'a' default-reduction
+ | 'a' default-reduction 'a'
+ | 'a' shift
+ ;
+
+default-reduction: /*empty*/ ;
+shift: 'b' ;
+
+// Provide another context in which all rules are useful so that this
+// test case looks a little more realistic.
+start: 'b' consistent-error-on-a-a 'c' ;
+]])
+
+m4_pushdef([AT_PREVIOUS_STATE_INPUT], [[a]])
+
+# Unfortunately, no expected tokens are reported even though 'b' can be
+# accepted. Nevertheless, the main point of this test is to make sure
+# that at least the unexpected token is reported. In a previous version
+# of Bison, it wasn't reported because the error is detected in a
+# consistent state with an error action, and that case always triggered
+# the simple "syntax error" message.
+#
+# The point isn't to test IELR here, but state merging happens to
+# complicate this example.
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type ielr]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[none]])
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type ielr
+ %glr-parser]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[none]])
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type ielr
+ %language "c++"]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[none]])
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type ielr
+ %language "java"]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[end of input]], [[none]])
+
+# Even canonical LR doesn't foresee the error for 'a'!
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type ielr
+ %define lr.default-reduction consistent]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[ab]])
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type ielr
+ %define lr.default-reduction accepting]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[ab]])
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type canonical-lr]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[ab]])
+
+# Only LAC gets it right.
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type canonical-lr
+ %define parse.lac full]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[b]])
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type ielr
+ %define parse.lac full]],
+ [AT_PREVIOUS_STATE_GRAMMAR],
+ [AT_PREVIOUS_STATE_INPUT],
+ [[$end]], [[b]])
+
+m4_popdef([AT_PREVIOUS_STATE_GRAMMAR])
+m4_popdef([AT_PREVIOUS_STATE_INPUT])
+
+m4_pushdef([AT_USER_ACTION_GRAMMAR],
+[[%nonassoc 'a';
+
+// If $$ = 0 here, then we know that the 'a' destructor is being invoked
+// incorrectly for the 'b' set in the semantic action below. All 'a'
+// tokens are returned by yylex, which sets $$ = 1.
+%destructor {
+ if (!$$)
+ fprintf (stderr, "Wrong destructor.\n");
+} 'a';
+
+// Rather than depend on an inconsistent state to induce reading a
+// lookahead as in the previous grammar, just assign the lookahead in a
+// semantic action. That lookahead isn't needed before either error
+// action is encountered. In a previous version of Bison, this was a
+// problem as it meant yychar was not translated into yytoken before
+// either error action. The second error action thus invoked a
+// destructor that it selected according to the incorrect yytoken. The
+// first error action would have reported an incorrect unexpected token
+// except that, due to the bug described in the previous grammar, the
+// unexpected token was not reported at all.
+start: error-reduce consistent-error 'a' { USE ($][3); } ;
+
+error-reduce:
+ 'a' 'a' consistent-reduction consistent-error 'a'
+ { USE (($][1, $][2, $][5)); }
+| 'a' error
+ { USE ($][1); }
+;
+
+consistent-reduction: /*empty*/ {
+ assert (yychar == YYEMPTY);
+ yylval = 0;
+ yychar = 'b';
+} ;
+
+consistent-error:
+ 'a' { USE ($][1); }
+| /*empty*/ %prec 'a'
+;
+
+// Provide another context in which all rules are useful so that this
+// test case looks a little more realistic.
+start: 'b' consistent-error 'b' ;
+]])
+m4_pushdef([AT_USER_ACTION_INPUT], [[aa]])
+
+AT_CONSISTENT_ERRORS_CHECK([[]],
+ [AT_USER_ACTION_GRAMMAR],
+ [AT_USER_ACTION_INPUT],
+ [['b']], [[none]])
+AT_CONSISTENT_ERRORS_CHECK([[%glr-parser]],
+ [AT_USER_ACTION_GRAMMAR],
+ [AT_USER_ACTION_INPUT],
+ [['b']], [[none]])
+# No C++ or Java test because yychar cannot be manipulated by users.
+
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.default-reduction consistent]],
+ [AT_USER_ACTION_GRAMMAR],
+ [AT_USER_ACTION_INPUT],
+ [['b']], [[none]])
+
+# Canonical LR doesn't foresee the error for 'a'!
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.default-reduction accepting]],
+ [AT_USER_ACTION_GRAMMAR],
+ [AT_USER_ACTION_INPUT],
+ [[$end]], [[a]])
+AT_CONSISTENT_ERRORS_CHECK([[%define lr.type canonical-lr]],
+ [AT_USER_ACTION_GRAMMAR],
+ [AT_USER_ACTION_INPUT],
+ [[$end]], [[a]])
+
+AT_CONSISTENT_ERRORS_CHECK([[%define parse.lac full]],
+ [AT_USER_ACTION_GRAMMAR],
+ [AT_USER_ACTION_INPUT],
+ [['b']], [[none]])
+AT_CONSISTENT_ERRORS_CHECK([[%define parse.lac full
+ %define lr.default-reduction accepting]],
+ [AT_USER_ACTION_GRAMMAR],
+ [AT_USER_ACTION_INPUT],
+ [[$end]], [[none]])
+
+m4_popdef([AT_USER_ACTION_GRAMMAR])
+m4_popdef([AT_USER_ACTION_INPUT])
+
+m4_popdef([AT_CONSISTENT_ERRORS_CHECK])
+
+
+
+
+## ------------------------------------------------------- ##
+## LAC: %nonassoc requires splitting canonical LR states. ##
+## ------------------------------------------------------- ##
+
+# This test case demonstrates that, when %nonassoc is used, canonical
+# LR(1) parser table construction followed by conflict resolution
+# without further state splitting is not always sufficient to produce a
+# parser that can detect all syntax errors as soon as possible on one
+# token of lookahead. However, LAC solves the problem completely even
+# with minimal LR parser tables.
+
+AT_SETUP([[LAC: %nonassoc requires splitting canonical LR states]])
+AT_BISON_OPTION_PUSHDEFS
+AT_DATA_GRAMMAR([[input.y]],
+[[%code {
+ #include <stdio.h>
+ ]AT_YYERROR_DECLARE[
+ ]AT_YYLEX_DECLARE[
+}
+
+%error-verbose
+%nonassoc 'a'
+
+%%
+
+start:
+ 'a' problem 'a' // First context.
+| 'b' problem 'b' // Second context.
+| 'c' reduce-nonassoc // Just makes reduce-nonassoc useful.
+;
+
+problem:
+ look reduce-nonassoc
+| look 'a'
+| look 'b'
+;
+
+// For the state reached after shifting the 'a' in these productions,
+// lookahead sets are the same in both the first and second contexts.
+// Thus, canonical LR reuses the same state for both contexts. However,
+// the lookahead 'a' for the reduction "look: 'a'" later becomes an
+// error action only in the first context. In order to immediately
+// detect the syntax error on 'a' here for only the first context, this
+// canonical LR state would have to be split into two states, and the
+// 'a' lookahead would have to be removed from only one of the states.
+look:
+ 'a' // Reduction lookahead set is always ['a', 'b'].
+| 'a' 'b'
+| 'a' 'c' // 'c' is forgotten as an expected token.
+;
+
+reduce-nonassoc: %prec 'a';
+
+%%
+]AT_YYERROR_DEFINE[
+]AT_YYLEX_DEFINE(["aaa"])[
+]AT_MAIN_DEFINE
+])
+AT_BISON_OPTION_POPDEFS
+
+# Show canonical LR's failure.
+AT_BISON_CHECK([[-Dlr.type=canonical-lr -o input.c input.y]],
+ [[0]], [[]],
+[[input.y: warning: 2 shift/reduce conflicts [-Wconflicts-sr]
+]])
+AT_COMPILE([[input]])
+AT_PARSER_CHECK([[./input]], [[1]], [[]],
+[[syntax error, unexpected 'a', expecting 'b'
+]])
+
+# It's corrected by LAC.
+AT_BISON_CHECK([[-Dlr.type=canonical-lr -Dparse.lac=full \
+ -o input.c input.y]], [[0]], [[]],
+[[input.y: warning: 2 shift/reduce conflicts [-Wconflicts-sr]
+]])
+AT_COMPILE([[input]])
+AT_PARSER_CHECK([[./input]], [[1]], [[]],
+[[syntax error, unexpected 'a', expecting 'b' or 'c'
+]])
+
+# IELR is sufficient when LAC is used.
+AT_BISON_CHECK([[-Dlr.type=ielr -Dparse.lac=full -o input.c input.y]],
+ [[0]], [[]],
+[[input.y: warning: 2 shift/reduce conflicts [-Wconflicts-sr]
+]])
+AT_COMPILE([[input]])
+AT_PARSER_CHECK([[./input]], [[1]], [[]],
+[[syntax error, unexpected 'a', expecting 'b' or 'c'
+]])
+
+AT_CLEANUP
+
projects / bison.git / blobdiff