X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/e0c40012bd7c94c88a0bb9886e37de9622729d9a..47e7600a21bf1d2be5ee2f2f337b57072e128c30:/tests/regression.at diff --git a/tests/regression.at b/tests/regression.at index 68f02adc..3f83ac46 100644 --- a/tests/regression.at +++ b/tests/regression.at @@ -1,5 +1,5 @@ # Bison Regressions. -*- Autotest -*- -# Copyright 2001 Free Software Foundation, Inc. +# Copyright (C) 2001, 2002, 2003 Free Software Foundation, Inc. # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by @@ -18,6 +18,67 @@ AT_BANNER([[Regression tests.]]) + +## ------------------------- ## +## Early token definitions. ## +## ------------------------- ## + + +AT_SETUP([Early token definitions]) + +# Found in GCJ: they expect the tokens to be defined before the user +# prologue, so that they can use the token definitions in it. + +AT_DATA_GRAMMAR([input.y], +[[%{ +void yyerror (const char *s); +int yylex (void); +%} + +%union +{ + int val; +}; +%{ +#ifndef MY_TOKEN +# error "MY_TOKEN not defined." +#endif +%} +%token MY_TOKEN +%% +exp: MY_TOKEN; +%% +]]) + +AT_CHECK([bison -o input.c input.y]) +AT_COMPILE([input.o], [-c input.c]) + +AT_CLEANUP + + + +## ---------------- ## +## Braces parsing. ## +## ---------------- ## + + +AT_SETUP([Braces parsing]) + +AT_DATA([input.y], +[[/* Bison used to swallow the character after `}'. */ + +%% +exp: { tests = {{{{{{{{{{}}}}}}}}}}; }; +%% +]]) + +AT_CHECK([bison -v -o input.c input.y]) + +AT_CHECK([grep 'tests = {{{{{{{{{{}}}}}}}}}};' input.c], 0, [ignore]) + +AT_CLEANUP + + ## ------------------ ## ## Duplicate string. ## ## ------------------ ## @@ -25,7 +86,7 @@ AT_BANNER([[Regression tests.]]) AT_SETUP([Duplicate string]) -AT_DATA([duplicate.y], +AT_DATA([input.y], [[/* `Bison -v' used to dump core when two tokens are defined with the same string, as LE and GE below. */ @@ -38,421 +99,696 @@ exp: '(' exp ')' | NUM ; %% ]]) -AT_CHECK([bison -v duplicate.y -o duplicate.c], 0, ignore, ignore) +AT_CHECK([bison -v -o input.c input.y], 0, [], +[[input.y:6.8-14: warning: symbol `"<="' used more than once as a literal string +]]) AT_CLEANUP -## ------------------------- ## -## Unresolved SR Conflicts. ## -## ------------------------- ## +## ------------------- ## +## Rule Line Numbers. ## +## ------------------- ## -AT_SETUP([Unresolved SR Conflicts]) +AT_SETUP([Rule Line Numbers]) + +AT_KEYWORDS([report]) AT_DATA([input.y], -[[%token NUM OP -%% -exp: exp OP exp | NUM; +[[%% +expr: +'a' + +{ + +} + +'b' + +{ + +} + +| + + +{ + + +} + +'c' + +{ + +}; ]]) -AT_CHECK([bison input.y -o input.c -v], 0, [], -[input.y contains 1 shift/reduce conflict. -]) +AT_CHECK([bison -o input.c -v input.y]) # Check the contents of the report. AT_CHECK([cat input.output], [], -[[State 4 contains 1 shift/reduce conflict. +[[Grammar + + 0 $accept: expr $end + + 1 @1: /* empty */ + + 2 expr: 'a' @1 'b' -Grammar + 3 @2: /* empty */ + + 4 expr: @2 'c' - Number, Line, Rule - 1 3 exp -> exp OP exp - 2 3 exp -> NUM Terminals, with rules where they appear -$ (-1) +$end (0) 0 +'a' (97) 2 +'b' (98) 2 +'c' (99) 4 error (256) -NUM (257) 2 -OP (258) 1 + Nonterminals, with rules where they appear -exp (5) - on left: 1 2, on right: 1 +$accept (6) + on left: 0 +expr (7) + on left: 2 4, on right: 0 +@1 (8) + on left: 1, on right: 2 +@2 (9) + on left: 3, on right: 4 state 0 - NUM shift, and go to state 1 + 0 $accept: . expr $end - exp go to state 2 + 'a' shift, and go to state 1 + $default reduce using rule 3 (@2) + + expr go to state 2 + @2 go to state 3 state 1 - exp -> NUM . (rule 2) + 2 expr: 'a' . @1 'b' - $default reduce using rule 2 (exp) + $default reduce using rule 1 (@1) + @1 go to state 4 state 2 - exp -> exp . OP exp (rule 1) - - $ go to state 5 - OP shift, and go to state 3 + 0 $accept: expr . $end + $end shift, and go to state 5 state 3 - exp -> exp OP . exp (rule 1) + 4 expr: @2 . 'c' - NUM shift, and go to state 1 + 'c' shift, and go to state 6 - exp go to state 4 +state 4 + 2 expr: 'a' @1 . 'b' -state 4 + 'b' shift, and go to state 7 - exp -> exp . OP exp (rule 1) - exp -> exp OP exp . (rule 1) - OP shift, and go to state 3 +state 5 - OP [reduce using rule 1 (exp)] - $default reduce using rule 1 (exp) + 0 $accept: expr $end . + $default accept -state 5 +state 6 - $ go to state 6 + 4 expr: @2 'c' . + $default reduce using rule 4 (expr) -state 6 +state 7 + + 2 expr: 'a' @1 'b' . - $default accept + $default reduce using rule 2 (expr) ]]) AT_CLEANUP -## --------------------- ## -## Solved SR Conflicts. ## -## --------------------- ## -AT_SETUP([Solved SR Conflicts]) +## ---------------------- ## +## Mixing %token styles. ## +## ---------------------- ## + +AT_SETUP([Mixing %token styles]) + +# Taken from the documentation. AT_DATA([input.y], -[[%token NUM OP -%right OP +[[%token OR "||" +%token LE 134 "<=" +%left OR "<=" +%% +exp: ; %% -exp: exp OP exp | NUM; ]]) -AT_CHECK([bison input.y -o input.c -v], 0, [], []) - -# Check the contents of the report. -AT_CHECK([cat input.output], [], -[[Conflict in state 4 between rule 1 and token OP resolved as shift. +AT_CHECK([bison -v -o input.c input.y]) -Grammar - - Number, Line, Rule - 1 4 exp -> exp OP exp - 2 4 exp -> NUM - -Terminals, with rules where they appear - -$ (-1) -error (256) -NUM (257) 2 -OP (258) 1 +AT_CLEANUP -Nonterminals, with rules where they appear -exp (5) - on left: 1 2, on right: 1 +## ---------------- ## +## Invalid inputs. ## +## ---------------- ## -state 0 - NUM shift, and go to state 1 +AT_SETUP([Invalid inputs]) - exp go to state 2 +AT_DATA([input.y], +[[%% +? +default: 'a' } +%& +%a-does-not-exist +%- +%{ +]]) +AT_CHECK([bison input.y], [1], [], +[[input.y:2.1: invalid character: `?' +input.y:3.14: invalid character: `}' +input.y:4.1: invalid character: `%' +input.y:4.2: invalid character: `&' +input.y:5.1-17: invalid directive: `%a-does-not-exist' +input.y:6.1: invalid character: `%' +input.y:6.2: invalid character: `-' +input.y:7.1-8.0: missing `%}' at end of file +]]) +AT_CLEANUP -state 1 - exp -> NUM . (rule 2) - $default reduce using rule 2 (exp) +## ------------------- ## +## Token definitions. ## +## ------------------- ## +AT_SETUP([Token definitions]) -state 2 +# Bison managed, when fed with `%token 'f' "f"' to #define 'f'! +AT_DATA_GRAMMAR([input.y], +[%{ +void yyerror (const char *s); +int yylex (void); +%} +[%token MYEOF 0 "end of file" +%token 'a' "a" +%token b "b" +%token c 'c' +%token 'd' d +%% +exp: "a"; +]]) - exp -> exp . OP exp (rule 1) +AT_CHECK([bison -o input.c input.y]) +AT_COMPILE([input.o], [-c input.c]) +AT_CLEANUP - $ go to state 5 - OP shift, and go to state 3 +## -------------------- ## +## Characters Escapes. ## +## -------------------- ## -state 3 - exp -> exp OP . exp (rule 1) +AT_SETUP([Characters Escapes]) - NUM shift, and go to state 1 +AT_DATA_GRAMMAR([input.y], +[%{ +void yyerror (const char *s); +int yylex (void); +%} +[%% +exp: + '\'' "\'" +| '\"' "\"" +| '"' "'" +; +]]) +# Pacify font-lock-mode: " - exp go to state 4 +AT_CHECK([bison -o input.c input.y]) +AT_COMPILE([input.o], [-c input.c]) +AT_CLEANUP -state 4 +## -------------- ## +## Web2c Report. ## +## -------------- ## - exp -> exp . OP exp (rule 1) - exp -> exp OP exp . (rule 1) +# The generation of the reduction was once wrong in Bison, and made it +# miss some reductions. In the following test case, the reduction on +# `undef_id_tok' in state 1 was missing. This is stripped down from +# the actual web2c.y. - OP shift, and go to state 3 +AT_SETUP([Web2c Report]) - $default reduce using rule 1 (exp) +AT_KEYWORDS([report]) +AT_DATA([input.y], +[[%token undef_id_tok const_id_tok +%start CONST_DEC_PART + +%% +CONST_DEC_PART: + CONST_DEC_LIST + ; + +CONST_DEC_LIST: + CONST_DEC + | CONST_DEC_LIST CONST_DEC + ; + +CONST_DEC: + { } undef_id_tok '=' const_id_tok ';' + ; +%% +]]) -state 5 +AT_CHECK([bison -v input.y]) +AT_CHECK([cat input.output], 0, +[[Grammar - $ go to state 6 + 0 $accept: CONST_DEC_PART $end + 1 CONST_DEC_PART: CONST_DEC_LIST + 2 CONST_DEC_LIST: CONST_DEC + 3 | CONST_DEC_LIST CONST_DEC -state 6 + 4 @1: /* empty */ - $default accept -]]) + 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';' -AT_CLEANUP +Terminals, with rules where they appear +$end (0) 0 +';' (59) 5 +'=' (61) 5 +error (256) +undef_id_tok (258) 5 +const_id_tok (259) 5 -## -------------------- ## -## %expect not enough. ## -## -------------------- ## -AT_SETUP([%expect not enough]) +Nonterminals, with rules where they appear -AT_DATA([input.y], -[[%token NUM OP -%expect 0 -%% -exp: exp OP exp | NUM; -]]) +$accept (7) + on left: 0 +CONST_DEC_PART (8) + on left: 1, on right: 0 +CONST_DEC_LIST (9) + on left: 2 3, on right: 1 3 +CONST_DEC (10) + on left: 5, on right: 2 3 +@1 (11) + on left: 4, on right: 5 -AT_CHECK([bison input.y -o input.c], 1, [], -[input.y contains 1 shift/reduce conflict. -expected 0 shift/reduce conflicts -]) -AT_CLEANUP +state 0 -## --------------- ## -## %expect right. ## -## --------------- ## + 0 $accept: . CONST_DEC_PART $end -AT_SETUP([%expect right]) + $default reduce using rule 4 (@1) -AT_DATA([input.y], -[[%token NUM OP -%expect 1 -%% -exp: exp OP exp | NUM; -]]) + CONST_DEC_PART go to state 1 + CONST_DEC_LIST go to state 2 + CONST_DEC go to state 3 + @1 go to state 4 -AT_CHECK([bison input.y -o input.c], 0) -AT_CLEANUP +state 1 -## ------------------ ## -## %expect too much. ## -## ------------------ ## + 0 $accept: CONST_DEC_PART . $end -AT_SETUP([%expect too much]) + $end shift, and go to state 5 -AT_DATA([input.y], -[[%token NUM OP -%expect 2 -%% -exp: exp OP exp | NUM; -]]) -AT_CHECK([bison input.y -o input.c], 1, [], -[input.y contains 1 shift/reduce conflict. -expected 2 shift/reduce conflicts -]) -AT_CLEANUP +state 2 + 1 CONST_DEC_PART: CONST_DEC_LIST . + 3 CONST_DEC_LIST: CONST_DEC_LIST . CONST_DEC -## ---------------------- ## -## Mixing %token styles. ## -## ---------------------- ## + undef_id_tok reduce using rule 4 (@1) + $default reduce using rule 1 (CONST_DEC_PART) + CONST_DEC go to state 6 + @1 go to state 4 -AT_SETUP([Mixing %token styles]) -# Taken from the documentation. -AT_DATA([input.y], -[[%token OR "||" -%token LE 134 "<=" -%left OR "<=" -%% -exp: ; -%% -]]) +state 3 -AT_CHECK([bison -v input.y -o input.c], 0, ignore, ignore) + 2 CONST_DEC_LIST: CONST_DEC . -AT_CLEANUP + $default reduce using rule 2 (CONST_DEC_LIST) +state 4 -## ---------------------- ## -## %union and --defines. ## -## ---------------------- ## + 5 CONST_DEC: @1 . undef_id_tok '=' const_id_tok ';' + undef_id_tok shift, and go to state 7 -AT_SETUP([%union and --defines]) -AT_DATA([union.y], -[%union -{ - int integer; - char *string ; -} -%% -exp: {}; -]) +state 5 -AT_CHECK([bison --defines union.y]) + 0 $accept: CONST_DEC_PART $end . -AT_CLEANUP + $default accept -## --------------------------------------- ## -## Duplicate '/' in C comments in %union ## -## --------------------------------------- ## +state 6 + 3 CONST_DEC_LIST: CONST_DEC_LIST CONST_DEC . -AT_SETUP([%union and C comments]) + $default reduce using rule 3 (CONST_DEC_LIST) -AT_DATA([union-comment.y], -[%union -{ - /* The int. */ int integer; - /* The string. */ char *string ; -} -%% -exp: {}; -]) -AT_CHECK([bison union-comment.y]) -AT_CHECK([fgrep '//*' union-comment.tab.c], [1], []) +state 7 -AT_CLEANUP + 5 CONST_DEC: @1 undef_id_tok . '=' const_id_tok ';' + '=' shift, and go to state 8 -## ----------------- ## -## Invalid input 1. ## -## ----------------- ## +state 8 -AT_SETUP([Invalid input: 1]) + 5 CONST_DEC: @1 undef_id_tok '=' . const_id_tok ';' -AT_DATA([input.y], -[[%% -? -]]) + const_id_tok shift, and go to state 9 -AT_CHECK([bison input.y], [1], [], -[[input.y:2: invalid input: `?' -input.y:3: fatal error: no rules in the input grammar -]]) -AT_CLEANUP +state 9 + 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok . ';' -## ----------------- ## -## Invalid input 2. ## -## ----------------- ## + ';' shift, and go to state 10 -AT_SETUP([Invalid input: 2]) +state 10 -AT_DATA([input.y], -[[%% -default: 'a' } -]]) + 5 CONST_DEC: @1 undef_id_tok '=' const_id_tok ';' . -AT_CHECK([bison input.y], [1], [], -[[input.y:2: invalid input: `}' + $default reduce using rule 5 (CONST_DEC) ]]) AT_CLEANUP +## --------------- ## +## Web2c Actions. ## +## --------------- ## + +# The generation of the mapping `state -> action' was once wrong in +# extremely specific situations. web2c.y exhibits this situation. +# Below is a stripped version of the grammar. It looks like one can +# simplify it further, but just don't: it is tuned to exhibit a bug, +# which disapears when applying sane grammar transformations. +# +# It used to be wrong on yydefact only: +# +# static const short yydefact[] = +# { +# - 2, 0, 1, 0, 0, 2, 3, 2, 5, 4, +# + 2, 0, 1, 0, 0, 0, 3, 2, 5, 4, +# 0, 0 +# }; +# +# but let's check all the tables. -## -------------------- ## -## Invalid %directive. ## -## -------------------- ## +AT_SETUP([Web2c Actions]) -AT_SETUP([Invalid %directive]) +AT_KEYWORDS([report]) AT_DATA([input.y], -[[%invalid +[[%% +statement: struct_stat; +struct_stat: /* empty. */ | if else; +if: "if" "const" "then" statement; +else: "else" statement; +%% ]]) -AT_CHECK([bison input.y], [1], [], -[[input.y:1: unrecognized: %invalid -input.y:1: Skipping to next % -input.y:2: fatal error: no input grammar +AT_CHECK([bison -v -o input.c input.y]) + +# Check only the tables. We don't use --no-parser, because it is +# still to be implemented in the experimental branch of Bison. +[sed -n 's/ *$//;/^static const.*\[\] =/,/^}/p' input.c >tables.c] + +AT_CHECK([[cat tables.c]], 0, +[[static const unsigned char yytranslate[] = +{ + 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, + 2, 2, 2, 2, 2, 2, 1, 2, 3, 4, + 5, 6 +}; +static const unsigned char yyprhs[] = +{ + 0, 0, 3, 5, 6, 9, 14 +}; +static const yysigned_char yyrhs[] = +{ + 8, 0, -1, 9, -1, -1, 10, 11, -1, 3, + 4, 5, 8, -1, 6, 8, -1 +}; +static const unsigned char yyrline[] = +{ + 0, 2, 2, 3, 3, 4, 5 +}; +static const char *const yytname[] = +{ + "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"", + "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0 +}; +static const unsigned short yytoknum[] = +{ + 0, 256, 257, 258, 259, 260, 261 +}; +static const unsigned char yyr1[] = +{ + 0, 7, 8, 9, 9, 10, 11 +}; +static const unsigned char yyr2[] = +{ + 0, 2, 1, 0, 2, 4, 2 +}; +static const unsigned char yydefact[] = +{ + 3, 0, 0, 2, 0, 0, 1, 3, 4, 3, + 6, 5 +}; +static const yysigned_char yydefgoto[] = +{ + -1, 2, 3, 4, 8 +}; +static const yysigned_char yypact[] = +{ + -2, -1, 4, -8, 0, 2, -8, -2, -8, -2, + -8, -8 +}; +static const yysigned_char yypgoto[] = +{ + -8, -7, -8, -8, -8 +}; +static const unsigned char yytable[] = +{ + 10, 1, 11, 5, 6, 0, 7, 9 +}; +static const yysigned_char yycheck[] = +{ + 7, 3, 9, 4, 0, -1, 6, 5 +}; +static const unsigned char yystos[] = +{ + 0, 3, 8, 9, 10, 4, 0, 6, 11, 5, + 8, 8 +}; ]]) AT_CLEANUP +## ------------------------- ## +## yycheck Bound Violation. ## +## ------------------------- ## -## --------------------- ## -## Invalid CPP headers. ## -## --------------------- ## -# AT_TEST_CPP_GUARD_H([INPUT-FILE-BASE) -# ------------------------------------- -m4_define([AT_TEST_CPP_GUARD_H], -[AT_SETUP([Invalid CPP guards: $1]) +# _AT_DATA_DANCER_Y(BISON-OPTIONS) +# -------------------------------- +# The following grammar, taken from Andrew Suffield's GPL'd implementation +# of DGMTP, the Dancer Generic Message Transport Protocol, used to violate +# yycheck's bounds where issuing a verbose error message. Keep this test +# so that possible bound checking compilers could check all the skeletons. +m4_define([_AT_DATA_DANCER_Y], +[AT_DATA_GRAMMAR([dancer.y], +[%{ +AT_LALR1_CC_IF( +[static int yylex (int *lval);], +[#include +static void yyerror (const char *s); +static int yylex ();]) +%} +$1 +%token ARROW INVALID NUMBER STRING DATA +%defines +%verbose +%error-verbose +/* Grammar follows */ +%% +line: header body + ; + +header: '<' from ARROW to '>' type ':' + | '<' ARROW to '>' type ':' + | ARROW to type ':' + | type ':' + | '<' '>' + ; + +from: DATA + | STRING + | INVALID + ; + +to: DATA + | STRING + | INVALID + ; + +type: DATA + | STRING + | INVALID + ; + +body: /* empty */ + | body member + ; + +member: STRING + | DATA + | '+' NUMBER + | '-' NUMBER + | NUMBER + | INVALID + ; +%% +AT_LALR1_CC_IF( +[/* Currently, print_ is required in C++. */ +void +yy::Parser::print_ () +{ +} -# Possibly create inner directories. -dirname=`AS_DIRNAME([$1])` -AS_MKDIR_P([$dirname]) +/* A C++ error reporting function. */ +void +yy::Parser::error_ () +{ + std::cerr << message << std::endl; +} -AT_DATA([$1.y], -[%% -dummy: +int +yyparse (void) +{ + yy::Parser parser (!!YYDEBUG); + return parser.parse (); +} +], +[static void +yyerror (const char *s) +{ + fprintf (stderr, "%s\n", s); +}]) + +static int +yylex (AT_LALR1_CC_IF([int *lval])) +[{ + static int toknum = 0; + int tokens[] = + { + ':', -1 + }; + return tokens[toknum++]; +}] + +int +main (void) +{ + return yyparse (); +} ]) - -AT_CHECK([bison --defines=$1.h $1.y]) - -# CPP should be happy with it. -AT_CHECK([$CC -E $1.h], 0, [ignore]) - +])# _AT_DATA_DANCER_Y + + +# AT_CHECK_DANCER(BISON-OPTIONS) +# ------------------------------ +# Generate the grammar, compile it, run it. +m4_define([AT_CHECK_DANCER], +[AT_SETUP([Dancer $1]) +AT_BISON_OPTION_PUSHDEFS([$1]) +_AT_DATA_DANCER_Y([$1]) +AT_CHECK([bison -o dancer.c dancer.y]) +AT_LALR1_CC_IF([AT_COMPILE_CXX([dancer])], + [AT_COMPILE([dancer])]) +AT_PARSER_CHECK([./dancer], 1, [], +[syntax error, unexpected ':' +]) +AT_BISON_OPTION_POPDEFS AT_CLEANUP ]) -AT_TEST_CPP_GUARD_H([input/input]) -AT_TEST_CPP_GUARD_H([9foo]) +AT_CHECK_DANCER() +AT_CHECK_DANCER([%glr-parser]) +AT_CHECK_DANCER([%skeleton "lalr1.cc"])