X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/68dbdee86acf38426291289a5e55b7f200fa207b..2a6b66c5fd063c8868f239a78ecfb11c899fd22c:/tests/regression.at diff --git a/tests/regression.at b/tests/regression.at index 01bbb7a3..8a285af1 100644 --- a/tests/regression.at +++ b/tests/regression.at @@ -1,7 +1,6 @@ # Bison Regressions. -*- Autotest -*- -# Copyright (C) 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008 Free Software -# Foundation, Inc. +# Copyright (C) 2001-2012 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 @@ -25,10 +24,11 @@ AT_BANNER([[Regression tests.]]) AT_SETUP([Trivial grammars]) +AT_BISON_OPTION_PUSHDEFS AT_DATA_GRAMMAR([input.y], [[%{ -void yyerror (char const *); -int yylex (void); +]AT_YYERROR_DECLARE_EXTERN[ +]AT_YYLEX_DECLARE_EXTERN[ #define YYSTYPE int * %} @@ -38,9 +38,10 @@ int yylex (void); program: 'x'; ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-o input.c input.y]) -AT_COMPILE([input.o], [-c input.c]) +AT_COMPILE([input.o]) AT_COMPILE([input.o], [-DYYDEBUG -c input.c]) AT_CLEANUP @@ -53,10 +54,11 @@ AT_CLEANUP AT_SETUP([YYSTYPE typedef]) +AT_BISON_OPTION_PUSHDEFS AT_DATA_GRAMMAR([input.y], [[%{ -void yyerror (char const *); -int yylex (void); +]AT_YYERROR_DECLARE_EXTERN[ +]AT_YYLEX_DECLARE_EXTERN[ typedef union { char const *val; } YYSTYPE; %} @@ -66,9 +68,10 @@ typedef union { char const *val; } YYSTYPE; program: { $$ = ""; }; ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-o input.c input.y]) -AT_COMPILE([input.o], [-c input.c]) +AT_COMPILE([input.o]) AT_CLEANUP @@ -84,10 +87,11 @@ AT_SETUP([Early token definitions with --yacc]) # 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_BISON_OPTION_PUSHDEFS AT_DATA_GRAMMAR([input.y], [[%{ -void yyerror (const char *s); -int yylex (void); +]AT_YYERROR_DECLARE_EXTERN[ +]AT_YYLEX_DECLARE_EXTERN[ %} %union @@ -104,9 +108,10 @@ int yylex (void); exp: MY_TOKEN; %% ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-y -o input.c input.y]) -AT_COMPILE([input.o], [-c input.c]) +AT_COMPILE([input.o]) AT_CLEANUP @@ -122,11 +127,12 @@ AT_SETUP([Early token definitions without --yacc]) # 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_BISON_OPTION_PUSHDEFS AT_DATA_GRAMMAR([input.y], [[%{ #include -void yyerror (const char *s); -int yylex (void); +]AT_YYERROR_DECLARE_EXTERN[ +]AT_YYLEX_DECLARE_EXTERN[ void print_my_token (void); %} @@ -147,9 +153,10 @@ print_my_token (void) exp: MY_TOKEN; %% ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-o input.c input.y]) -AT_COMPILE([input.o], [-c input.c]) +AT_COMPILE([input.o]) AT_CLEANUP @@ -162,13 +169,15 @@ AT_CLEANUP AT_SETUP([Braces parsing]) +AT_BISON_OPTION_PUSHDEFS AT_DATA([input.y], -[[/* Bison used to swallow the character after `}'. */ +[[/* Bison used to swallow the character after '}'. */ %% exp: { tests = {{{{{{{{{{}}}}}}}}}}; }; %% ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-v -o input.c input.y]) @@ -184,8 +193,9 @@ AT_CLEANUP AT_SETUP([Duplicate string]) +AT_BISON_OPTION_PUSHDEFS AT_DATA([input.y], -[[/* `Bison -v' used to dump core when two tokens are defined with the same +[[/* 'Bison -v' used to dump core when two tokens are defined with the same string, as LE and GE below. */ %token NUM @@ -196,9 +206,10 @@ AT_DATA([input.y], exp: '(' exp ')' | NUM ; %% ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-v -o input.c input.y], 0, [], -[[input.y:6.8-14: warning: symbol `"<="' used more than once as a literal string +[[input.y:6.8-14: warning: symbol "<=" used more than once as a literal string [-Wother] ]]) AT_CLEANUP @@ -212,6 +223,7 @@ AT_SETUP([Rule Line Numbers]) AT_KEYWORDS([report]) +AT_BISON_OPTION_PUSHDEFS AT_DATA([input.y], [[%% expr: @@ -241,6 +253,7 @@ expr: }; ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-o input.c -v input.y]) @@ -388,15 +401,15 @@ default: 'a' } ]]) AT_BISON_CHECK([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 -input.y:7.1-8.0: syntax error, unexpected %{...%} +[[input.y:2.1: error: invalid character: '?' +input.y:3.14: error: invalid character: '}' +input.y:4.1: error: invalid character: '%' +input.y:4.2: error: invalid character: '&' +input.y:5.1-17: error: invalid directive: '%a-does-not-exist' +input.y:6.1: error: invalid character: '%' +input.y:6.2: error: invalid character: '-' +input.y:7.1-8.0: error: missing '%}' at end of file +input.y:7.1-8.0: error: syntax error, unexpected %{...%} ]]) AT_CLEANUP @@ -415,7 +428,7 @@ AT_DATA([input.y], ]]) AT_BISON_CHECK([input.y], [1], [], -[[input.y:3.1-15: syntax error, unexpected %initial-action, expecting {...} +[[input.y:3.1-15: error: syntax error, unexpected %initial-action, expecting {...} ]]) AT_CLEANUP @@ -429,13 +442,14 @@ AT_CLEANUP AT_SETUP([Token definitions]) -# Bison managed, when fed with `%token 'f' "f"' to #define 'f'! +AT_BISON_OPTION_PUSHDEFS +# Bison managed, when fed with '%token 'f' "f"' to #define 'f'! AT_DATA_GRAMMAR([input.y], [%{ #include #include -void yyerror (const char *s); -int yylex (void); +]AT_YYERROR_DECLARE[ +]AT_YYLEX_DECLARE[ %} [%error-verbose %token MYEOF 0 "end of file" @@ -448,20 +462,8 @@ int yylex (void); %% exp: "a" "\\\'\?\"\a\b\f\n\r\t\v\001\201\x001\x000081??!"; %% -void -yyerror (char const *s) -{ - fprintf (stderr, "%s\n", s); -} - -int -yylex (void) -{ - static int called; - if (called++) - abort (); - return SPECIAL; -} +]AT_YYERROR_DEFINE[ +]AT_YYLEX_DEFINE([{ SPECIAL }])[ int main (void) @@ -469,14 +471,15 @@ main (void) return yyparse (); } ]]) +AT_BISON_OPTION_POPDEFS # Checking the warning message guarantees that the trigraph "??!" isn't # unnecessarily escaped here even though it would need to be if encoded in a # C-string literal. Also notice that unnecessary escaping, such as "\?", from # the user specification is eliminated. AT_BISON_CHECK([-o input.c input.y], [[0]], [[]], -[[input.y:22.8-14: warning: symbol SPECIAL redeclared -input.y:22.8-63: warning: symbol `"\\'?\"\a\b\f\n\r\t\v\001\201\001\201??!"' used more than once as a literal string +[[input.y:22.8-14: warning: symbol SPECIAL redeclared [-Wother] +input.y:22.8-63: warning: symbol "\\'?\"\a\b\f\n\r\t\v\001\201\001\201??!" used more than once as a literal string [-Wother] ]]) AT_COMPILE([input]) @@ -499,22 +502,24 @@ AT_CLEANUP AT_SETUP([Characters Escapes]) +AT_BISON_OPTION_PUSHDEFS AT_DATA_GRAMMAR([input.y], [%{ -void yyerror (const char *s); -int yylex (void); +]AT_YYERROR_DECLARE_EXTERN[ +]AT_YYLEX_DECLARE_EXTERN[ %} [%% exp: '\'' "\'" | '\"' "\"" -| '"' "'" +| '"' "'" /* Pacify font-lock-mode: ". */ ; ]]) -# Pacify font-lock-mode: " + +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-o input.c input.y]) -AT_COMPILE([input.o], [-c input.c]) +AT_COMPILE([input.o]) AT_CLEANUP @@ -525,7 +530,7 @@ AT_CLEANUP # 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 +# 'undef_id_tok' in state 1 was missing. This is stripped down from # the actual web2c.y. AT_SETUP([Web2c Report]) @@ -533,7 +538,7 @@ AT_SETUP([Web2c Report]) AT_KEYWORDS([report]) AT_DATA([input.y], -[[%token undef_id_tok const_id_tok +[[%token undef_id_tok const_id_tok %start CONST_DEC_PART @@ -543,12 +548,12 @@ CONST_DEC_PART: ; CONST_DEC_LIST: - CONST_DEC + CONST_DEC | CONST_DEC_LIST CONST_DEC ; CONST_DEC: - { } undef_id_tok '=' const_id_tok ';' + { } undef_id_tok '=' const_id_tok ';' ; %% ]]) @@ -687,7 +692,7 @@ AT_CLEANUP ## Web2c Actions. ## ## --------------- ## -# The generation of the mapping `state -> action' was once wrong in +# 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, @@ -761,38 +766,30 @@ static const yytype_uint8 yyrline[] = static const char *const yytname[] = { "$end", "error", "$undefined", "\"if\"", "\"const\"", "\"then\"", - "\"else\"", "$accept", "statement", "struct_stat", "if", "else", 0 + "\"else\"", "$accept", "statement", "struct_stat", "if", "else", YY_NULL }; static const yytype_uint16 yytoknum[] = { 0, 256, 257, 258, 259, 260, 261 }; -static const yytype_uint8 yyr1[] = -{ - 0, 7, 8, 9, 9, 10, 11 -}; -static const yytype_uint8 yyr2[] = +static const yytype_int8 yypact[] = { - 0, 2, 1, 0, 2, 4, 2 + -2, -1, 4, -8, 0, 2, -8, -2, -8, -2, + -8, -8 }; static const yytype_uint8 yydefact[] = { 3, 0, 0, 2, 0, 0, 1, 3, 4, 3, 6, 5 }; -static const yytype_int8 yydefgoto[] = -{ - -1, 2, 3, 4, 8 -}; -static const yytype_int8 yypact[] = -{ - -2, -1, 4, -8, 0, 2, -8, -2, -8, -2, - -8, -8 -}; static const yytype_int8 yypgoto[] = { -8, -7, -8, -8, -8 }; +static const yytype_int8 yydefgoto[] = +{ + -1, 2, 3, 4, 8 +}; static const yytype_uint8 yytable[] = { 10, 1, 11, 5, 6, 0, 7, 9 @@ -806,6 +803,14 @@ static const yytype_uint8 yystos[] = 0, 3, 8, 9, 10, 4, 0, 6, 11, 5, 8, 8 }; +static const yytype_uint8 yyr1[] = +{ + 0, 7, 8, 9, 9, 10, 11 +}; +static const yytype_uint8 yyr2[] = +{ + 0, 2, 1, 0, 2, 4, 2 +}; ]]) AT_CLEANUP @@ -824,13 +829,11 @@ AT_CLEANUP # so that possible bound checking compilers could check all the skeletons. m4_define([_AT_DATA_DANCER_Y], [AT_DATA_GRAMMAR([dancer.y], -[%{ -static int yylex (AT_LALR1_CC_IF([int *], [void])); -AT_LALR1_CC_IF([], -[#include -#include -static void yyerror (const char *);]) -%} +[[%code provides +{ + ]AT_YYERROR_DECLARE[ + ]AT_YYLEX_DECLARE[ +} $1 %token ARROW INVALID NUMBER STRING DATA %defines @@ -875,15 +878,10 @@ member: STRING | INVALID ; %% -AT_LALR1_CC_IF( -[/* A C++ error reporting function. */ -void -yy::parser::error (const std::string& m) -{ - std::cerr << m << std::endl; -} - -int +]AT_YYERROR_DEFINE[ +]AT_YYLEX_DEFINE([":"])[ +]AT_LALR1_CC_IF( +[int yyparse () { yy::parser parser; @@ -892,33 +890,14 @@ yyparse () #endif return parser.parse (); } -], -[static void -yyerror (const char *s) -{ - fprintf (stderr, "%s\n", s); -}]) - -static int -yylex (AT_LALR1_CC_IF([int *lval], [void])) -[{ - static int const tokens[] = - { - ':', -1 - }; - static size_t toknum; - ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC. */])[ - if (! (toknum < sizeof tokens / sizeof *tokens)) - abort (); - return tokens[toknum++]; -}] +])[ int main (void) { return yyparse (); } -]) +]]) ])# _AT_DATA_DANCER_Y @@ -929,7 +908,6 @@ m4_define([AT_CHECK_DANCER], [AT_SETUP([Dancer $1]) AT_BISON_OPTION_PUSHDEFS([$1]) _AT_DATA_DANCER_Y([$1]) -AT_BISON_CHECK([-o dancer.c dancer.y]) AT_FULL_COMPILE([dancer]) AT_PARSER_CHECK([./dancer], 1, [], [syntax error, unexpected ':' @@ -954,10 +932,10 @@ m4_define([_AT_DATA_EXPECT2_Y], [AT_DATA_GRAMMAR([expect2.y], [%{ static int yylex (AT_LALR1_CC_IF([int *], [void])); -AT_LALR1_CC_IF([], -[#include -#include -static void yyerror (const char *);]) +AT_LALR1_CC_IF([[#include ]], +[[#include +#include +]AT_YYERROR_DECLARE])[ %} $1 %defines @@ -974,58 +952,46 @@ e: e '+' t | t; t: A | B; %% -AT_LALR1_CC_IF( -[/* A C++ error reporting function. */ -void -yy::parser::error (const std::string& m) -{ - std::cerr << m << std::endl; -} - -int +]AT_YYERROR_DEFINE[ +]AT_LALR1_CC_IF( +[int yyparse () { yy::parser parser; return parser.parse (); } -], -[static void -yyerror (const char *s) -{ - fprintf (stderr, "%s\n", s); -}]) +])[ +#include static int -yylex (AT_LALR1_CC_IF([int *lval], [void])) -[{ +yylex (]AT_LALR1_CC_IF([int *lval], [void])[) +{ static int const tokens[] = { 1000, '+', '+', -1 }; static size_t toknum; ]AT_LALR1_CC_IF([*lval = 0; /* Pacify GCC. */])[ - if (! (toknum < sizeof tokens / sizeof *tokens)) - abort (); + assert (toknum < sizeof tokens / sizeof *tokens); return tokens[toknum++]; -}] +} int main (void) { return yyparse (); } -]) +]]) ])# _AT_DATA_EXPECT2_Y # AT_CHECK_EXPECT2(BISON-OPTIONS) -# ------------------------------ +# ------------------------------- # Generate the grammar, compile it, run it. m4_define([AT_CHECK_EXPECT2], [AT_SETUP([Expecting two tokens $1]) AT_BISON_OPTION_PUSHDEFS([$1]) _AT_DATA_EXPECT2_Y([$1]) -AT_BISON_CHECK([-o expect2.c expect2.y]) AT_FULL_COMPILE([expect2]) AT_PARSER_CHECK([./expect2], 1, [], [syntax error, unexpected '+', expecting A or B @@ -1048,12 +1014,12 @@ AT_SETUP([Braced code in declaration in rules section]) # Bison once mistook braced code in a declaration in the rules section to be a # rule action. - +AT_BISON_OPTION_PUSHDEFS AT_DATA_GRAMMAR([input.y], [[%{ #include -static void yyerror (char const *msg); -static int yylex (void); +]AT_YYERROR_DECLARE[ +]AT_YYLEX_DECLARE[ %} %error-verbose @@ -1063,7 +1029,7 @@ static int yylex (void); start: { printf ("Bison would once convert this action to a midrule because of the" - " subsequent braced code.\n"); + " subsequent braced code.\n"); } ; @@ -1071,18 +1037,8 @@ start: %printer { fprintf (yyoutput, "PRINTER"); } 'a'; %% - -static void -yyerror (char const *msg) -{ - fprintf (stderr, "%s\n", msg); -} - -static int -yylex (void) -{ - return 'a'; -} +]AT_YYERROR_DEFINE[ +]AT_YYLEX_DEFINE(["a"])[ int main (void) @@ -1091,6 +1047,7 @@ main (void) return !yyparse (); } ]]) +AT_BISON_OPTION_POPDEFS AT_BISON_CHECK([-t -o input.c input.y]) AT_COMPILE([input]) @@ -1149,7 +1106,7 @@ AT_SETUP([[Extra lookahead sets in report]]) # state's one item that (1) is associated with the same rule as the reduction # and (2) has its dot at the end of its RHS. Previously, Bison also # erroneously printed the lookahead set next to all of the state's other items -# associated with the same rule. This bug affected only the `.output' file and +# associated with the same rule. This bug affected only the '.output' file and # not the generated parser source code. AT_DATA([[input.y]], @@ -1188,21 +1145,24 @@ AT_SETUP([[Token number in precedence declaration]]) # POSIX says token numbers can be declared in %left, %right, and %nonassoc, but # we lost this in Bison 1.50. - +AT_BISON_OPTION_PUSHDEFS AT_DATA_GRAMMAR([input.y], [[%{ #include - void yyerror (char const *); - int yylex (void); + ]AT_YYERROR_DECLARE[ + ]AT_YYLEX_DECLARE[ %} %error-verbose +%right END 0 %left TK1 1 TK2 2 "tok alias" 3 %% -start: TK1 sr_conflict "tok alias" ; - +start: + TK1 sr_conflict "tok alias" + | start %prec END + ; sr_conflict: TK2 | TK2 "tok alias" @@ -1210,55 +1170,493 @@ sr_conflict: %% -void -yyerror (char const *msg) +]AT_YYERROR_DEFINE[ +]AT_YYLEX_DEFINE([{ 1, 2, 3, 0 }])[ + +int +main (void) +{ + return yyparse (); +} +]]) +AT_BISON_OPTION_POPDEFS + +AT_BISON_CHECK([[-o input.c input.y]], [[0]],, +[[input.y:23.5-19: warning: rule useless in parser due to conflicts: start: start [-Wother] +input.y:27.5-19: warning: rule useless in parser due to conflicts: sr_conflict: TK2 "tok alias" [-Wother] +]]) +AT_COMPILE([[input]]) +AT_PARSER_CHECK([[./input]]) + +AT_CLEANUP + + + +## --------------------------- ## +## parse-gram.y: LALR = IELR. ## +## --------------------------- ## + +# If parse-gram.y's LALR and IELR parser tables ever begin to differ, we +# need to fix parse-gram.y or start using IELR. + +AT_SETUP([[parse-gram.y: LALR = IELR]]) + +# Avoid tests/bison's dark magic by processing a local copy of the +# grammar. Avoid differences in synclines by telling bison that the +# output files have the same name. +[cp $abs_top_srcdir/src/parse-gram.y input.y] +AT_BISON_CHECK([[-o input.c -Dlr.type=lalr input.y]]) +[mv input.c lalr.c] +AT_CAPTURE_FILE([lalr.c]) +AT_BISON_CHECK([[-o input.c -Dlr.type=ielr input.y]]) +[mv input.c ielr.c] +AT_CAPTURE_FILE([ielr.c]) +AT_CHECK([[diff lalr.c ielr.c]], [[0]]) + +AT_CLEANUP + + + +## -------------------------------------------- ## +## parse.error=verbose and YYSTACK_USE_ALLOCA. ## +## -------------------------------------------- ## + +AT_SETUP([[parse.error=verbose and YYSTACK_USE_ALLOCA]]) + +AT_BISON_OPTION_PUSHDEFS +AT_DATA_GRAMMAR([input.y], +[[%code { + #include + ]AT_YYERROR_DECLARE[ + ]AT_YYLEX_DECLARE[ + #define YYSTACK_USE_ALLOCA 1 +} + +%define parse.error verbose + +%% + +start: check syntax_error syntax_error ; + +check: { - fprintf (stderr, "%s\n", msg); + if (128 < sizeof yymsgbuf) + { + fprintf (stderr, + "The initial size of yymsgbuf in yyparse has increased\n" + "since this test group was last updated. As a result,\n" + "this test group may no longer manage to induce a\n" + "reallocation of the syntax error message buffer.\n" + "This test group must be adjusted to produce a longer\n" + "error message.\n"); + YYABORT; + } } +; +// Induce a syntax error message whose total length is more than +// sizeof yymsgbuf in yyparse. Each token here is 64 bytes. +syntax_error: + "123456789112345678921234567893123456789412345678951234567896123A" +| "123456789112345678921234567893123456789412345678951234567896123B" +| error 'a' 'b' 'c' +; + +%% + +]AT_YYERROR_DEFINE[ + /* Induce two syntax error messages (which requires full error + recovery by shifting 3 tokens) in order to detect any loss of the + reallocated buffer. */ +]AT_YYLEX_DEFINE(["abc"])[ int -yylex (void) +main (void) { - static int const input[] = { 1, 2, 3, 0 }; - static int const *inputp = input; - return *inputp++; + return yyparse (); } +]]) +AT_BISON_OPTION_POPDEFS + +AT_BISON_CHECK([[-o input.c input.y]]) +AT_COMPILE([[input]]) +AT_PARSER_CHECK([[./input]], [[1]], [], +[[syntax error, unexpected 'a', expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B +syntax error, unexpected $end, expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B +]]) + +AT_CLEANUP + + +## ------------------------------ ## +## parse.error=verbose overflow. ## +## ------------------------------ ## + +# Imagine the case where YYSTACK_ALLOC_MAXIMUM = YYSIZE_MAXIMUM and an +# invocation of yysyntax_error has caused yymsg_alloc to grow to exactly +# YYSTACK_ALLOC_MAXIMUM (perhaps because the normal doubling of size had +# to be clipped to YYSTACK_ALLOC_MAXIMUM). In an old version of yacc.c, +# a subsequent invocation of yysyntax_error that overflows during its +# size calculation would return YYSIZE_MAXIMUM to yyparse. Then, +# yyparse would invoke yyerror using the old contents of yymsg. + +AT_SETUP([[parse.error=verbose overflow]]) + +AT_BISON_OPTION_PUSHDEFS +AT_DATA_GRAMMAR([input.y], +[[%code { + #include + ]AT_YYERROR_DECLARE[ + ]AT_YYLEX_DECLARE[ + + /* This prevents this test case from having to induce error messages + large enough to overflow size_t. */ + #define YYSIZE_T unsigned char + + /* Bring in malloc and set EXIT_SUCCESS so yacc.c doesn't try to + provide a malloc prototype using our YYSIZE_T. */ + #include + #ifndef EXIT_SUCCESS + # define EXIT_SUCCESS 0 + #endif + + /* Max depth is usually much smaller than YYSTACK_ALLOC_MAXIMUM, and + we don't want gcc to warn everywhere this constant would be too big + to make sense for our YYSIZE_T. */ + #define YYMAXDEPTH 100 +} + +%define parse.error verbose + +%% + +start: syntax_error1 check syntax_error2 ; + +// Induce a syntax error message whose total length causes yymsg in +// yyparse to be reallocated to size YYSTACK_ALLOC_MAXIMUM, which +// should be 255. Each token here is 64 bytes. +syntax_error1: + "123456789112345678921234567893123456789412345678951234567896123A" +| "123456789112345678921234567893123456789412345678951234567896123B" +| "123456789112345678921234567893123456789412345678951234567896123C" +| error 'a' 'b' 'c' +; + +check: +{ + if (yymsg_alloc != YYSTACK_ALLOC_MAXIMUM + || YYSTACK_ALLOC_MAXIMUM != YYSIZE_MAXIMUM + || YYSIZE_MAXIMUM != 255) + { + fprintf (stderr, + "The assumptions of this test group are no longer\n" + "valid, so it may no longer catch the error it was\n" + "designed to catch. Specifically, the following\n" + "values should all be 255:\n\n"); + fprintf (stderr, " yymsg_alloc = %d\n", yymsg_alloc); + fprintf (stderr, " YYSTACK_ALLOC_MAXIMUM = %d\n", + YYSTACK_ALLOC_MAXIMUM); + fprintf (stderr, " YYSIZE_MAXIMUM = %d\n", YYSIZE_MAXIMUM); + YYABORT; + } +} +; + +// Now overflow. +syntax_error2: + "123456789112345678921234567893123456789412345678951234567896123A" +| "123456789112345678921234567893123456789412345678951234567896123B" +| "123456789112345678921234567893123456789412345678951234567896123C" +| "123456789112345678921234567893123456789412345678951234567896123D" +| "123456789112345678921234567893123456789412345678951234567896123E" +; + +%% + +]AT_YYERROR_DEFINE[ + /* Induce two syntax error messages (which requires full error + recovery by shifting 3 tokens). */ +]AT_YYLEX_DEFINE(["abc"])[ int main (void) { + /* Push parsers throw away the message buffer between tokens, so skip + this test under maintainer-push-check. */ + if (YYPUSH) + return 77; return yyparse (); } ]]) -AT_BISON_CHECK([[-o input.c input.y]], [[0]],, -[[input.y:24.5-19: warning: rule useless in parser due to conflicts: sr_conflict: TK2 "tok alias" +AT_BISON_CHECK([[-o input.c input.y]]) + +# gcc warns about tautologies and fallacies involving comparisons for +# unsigned char. However, it doesn't produce these same warnings for +# size_t and many other types when the warnings would seem to make just +# as much sense. We ignore the warnings. +[CFLAGS="$NO_WERROR_CFLAGS"] +AT_COMPILE([[input]]) + +AT_PARSER_CHECK([[./input]], [[2]], [], +[[syntax error, unexpected 'a', expecting 123456789112345678921234567893123456789412345678951234567896123A or 123456789112345678921234567893123456789412345678951234567896123B or 123456789112345678921234567893123456789412345678951234567896123C +syntax error +memory exhausted +]]) +AT_BISON_OPTION_POPDEFS +AT_CLEANUP + + + +## ------------------------ ## +## LAC: Exploratory stack. ## +## ------------------------ ## + +AT_SETUP([[LAC: Exploratory stack]]) + +m4_pushdef([AT_LAC_CHECK], [ + +AT_BISON_OPTION_PUSHDEFS([$1]) + +AT_DATA_GRAMMAR([input.y], +[[%code { + #include + ]AT_YYERROR_DECLARE[ + int yylex (]AT_PURE_IF([[YYSTYPE *]], [[void]])[); +} + +]$1[ +%define parse.error verbose +%token 'c' + +%% + +// default reductions in inconsistent states +// v v v v v v v v v v v v v v +S: A B A A B A A A A B A A A A A A A B C C A A A A A A A A A A A A B ; +// ^ ^ ^ +// LAC reallocs + +A: 'a' | /*empty*/ { printf ("inconsistent default reduction\n"); } ; +B: 'b' ; +C: /*empty*/ { printf ("consistent default reduction\n"); } ; + +%% +]AT_YYERROR_DEFINE[ +int +yylex (]AT_PURE_IF([[YYSTYPE *v]], [[void]])[) +{ + static char const *input = "bbbbc";]AT_PURE_IF([[ + *v = 0;]])[ + return *input++; +} + +int +main (void) +{ + yydebug = 1; + return yyparse (); +} +]]) + +AT_BISON_CHECK([[-Dparse.lac=full -Dparse.lac.es-capacity-initial=1 \ + -Dparse.lac.memory-trace=full \ + -t -o input.c input.y]], [[0]], [], +[[input.y: warning: 21 shift/reduce conflicts [-Wconflicts-sr] ]]) AT_COMPILE([[input]]) -AT_PARSER_CHECK([[./input]]) +AT_PARSER_CHECK([[./input > stdout.txt 2> stderr.txt]], [[1]]) + +# Make sure syntax error doesn't forget that 'a' is expected. It would +# be forgotten without lookahead correction. +AT_CHECK([[grep 'syntax error,' stderr.txt]], [[0]], +[[syntax error, unexpected 'c', expecting 'a' or 'b' +]]) + +# Check number of default reductions in inconsistent states to be sure +# syntax error is detected before unnecessary reductions are performed. +AT_CHECK([[$PERL -0777 -ne 'print s/inconsistent default reduction//g;' \ + < stdout.txt || exit 77]], [[0]], [[14]]) + +# Check number of default reductions in consistent states to be sure +# it is performed before the syntax error is detected. +AT_CHECK([[$PERL -0777 -ne 'print s/\bconsistent default reduction//g;' \ + < stdout.txt || exit 77]], [[0]], [[2]]) + +# Check number of reallocs to be sure reallocated memory isn't somehow +# lost between LAC invocations. +AT_CHECK([[$PERL -0777 -ne 'print s/\(realloc//g;' < stderr.txt \ + || exit 77]], [[0]], [[3]]) + +AT_BISON_OPTION_POPDEFS +]) + +AT_LAC_CHECK([[%define api.push-pull pull]]) +AT_LAC_CHECK([[%define api.push-pull pull %define api.pure]]) +AT_LAC_CHECK([[%define api.push-pull both]]) +AT_LAC_CHECK([[%define api.push-pull both %define api.pure]]) + +m4_popdef([AT_LAC_CHECK]) AT_CLEANUP -## ----------------------------------------------- ## -## Fix user actions without a trailing semicolon. ## -## ----------------------------------------------- ## +## ------------------------ ## +## LAC: Memory exhaustion. ## +## ------------------------ ## -AT_SETUP([[Fix user actions without a trailing semicolon]]) +AT_SETUP([[LAC: Memory exhaustion]]) -# This feature is undocumented, but we accidentally broke it in 2.3a, and there -# was a complaint at: -# . +m4_pushdef([AT_LAC_CHECK], +[AT_BISON_OPTION_PUSHDEFS +AT_DATA_GRAMMAR([input.y], +[[%code { + #include + ]AT_YYERROR_DECLARE[ + ]AT_YYLEX_DECLARE[ + #define YYMAXDEPTH 8 +} -AT_DATA([input.y], -[[%% -start: {asdffdsa} ; +%error-verbose + +%% + +S: A A A A A A A A A ; +A: /*empty*/ | 'a' ; + +%% +]AT_YYERROR_DEFINE[ +]AT_YYLEX_DEFINE(["$1"])[ +int +main (void) +{ + yydebug = 1; + return yyparse (); +} ]]) -AT_BISON_CHECK([[-o input.c input.y]]) -AT_CHECK([[sed -n '/asdffdsa/s/^ *//p' input.c]], [[0]], -[[{asdffdsa;} +AT_BISON_CHECK([[-Dparse.lac=full -Dparse.lac.es-capacity-initial=1 \ + -t -o input.c input.y]], [[0]], [], +[[input.y: warning: 8 shift/reduce conflicts [-Wconflicts-sr] ]]) +AT_COMPILE([[input]]) +AT_BISON_OPTION_POPDEFS +]) + +# Check for memory exhaustion during parsing. +AT_LAC_CHECK([]) +AT_PARSER_CHECK([[./input]], [[2]], [], +[[Starting parse +Entering state 0 +Reading a token: Now at end of input. +LAC: initial context established for $end +LAC: checking lookahead $end: R2 G3 R2 G5 R2 G6 R2 G7 R2 G8 R2 G9 R2 G10 R2 G11 R2 (max size exceeded) +memory exhausted +Cleanup: discarding lookahead token $end () +Stack now 0 +]]) + +# Induce an immediate syntax error with an undefined token, and check +# for memory exhaustion while building syntax error message. +AT_LAC_CHECK([z], [[0]]) +AT_PARSER_CHECK([[./input]], [[2]], [], +[[Starting parse +Entering state 0 +Reading a token: Next token is token $undefined () +LAC: initial context established for $undefined +LAC: checking lookahead $undefined: Always Err +Constructing syntax error message +LAC: checking lookahead $end: R2 G3 R2 G5 R2 G6 R2 G7 R2 G8 R2 G9 R2 G10 R2 G11 R2 (max size exceeded) +syntax error +memory exhausted +Cleanup: discarding lookahead token $undefined () +Stack now 0 +]]) + +m4_popdef([AT_LAC_CHECK]) AT_CLEANUP + + +## ---------------------- ## +## Lex and parse params. ## +## ---------------------- ## + +# AT_TEST(SKELETON) +# ----------------- +# Check that the identifier of the params is properly fetched +# even when there are trailing blanks. + +m4_pushdef([AT_TEST], +[AT_SETUP([[Lex and parse params: $1]]) + +AT_BISON_OPTION_PUSHDEFS([%locations %skeleton $1]) + +## FIXME: Improve parsing of parse-param and use the generated +## yyerror. +AT_DATA_GRAMMAR([input.y], +[[%defines +%locations +%skeleton $1 +%union { int ival; } +%parse-param { int x } +// Spaces, tabs, and new lines. +%parse-param { @&t@ + int y @&t@ + @&t@ + @&t@ +} + +%{ +#include +#include + +]AT_SKEL_CC_IF([], [[ +static +void +yyerror (int x, int y, const char *msg) +{ + fprintf (stderr, "x: %d, y: %d, %s\n", x, y, msg); +}]])[ + + ]AT_YYLEX_DECLARE[ +%} + +%% +exp: 'a' { fprintf (stdout, "x: %d, y: %d\n", x, y); }; +%% +]AT_YYLEX_DEFINE(["a"])[ + +]AT_SKEL_CC_IF( +[AT_YYERROR_DEFINE + +int +yyparse (int x, int y) +{ + yy::parser parser(x, y); + return parser.parse (); +} +])[ + +int +main (void) +{ + return !!yyparse(1, 2); +} +]]) + +AT_FULL_COMPILE([input]) +AT_CHECK([./input], 0, [[x: 1, y: 2 +]]) +AT_BISON_OPTION_POPDEFS + +AT_CLEANUP +]) + +## FIXME: test Java, and iterate over skeletons. +AT_TEST("yacc.c") +AT_TEST("glr.c") +AT_TEST("lalr1.cc") +AT_TEST("glr.cc") + +m4_popdef([AT_TEST])