X-Git-Url: https://git.saurik.com/bison.git/blobdiff_plain/e8cd1ad655bcc704b06fb2f191dc3ac1df32b796..737406a32c201471699bfa0843d1f432f3ec29ab:/tests/regression.at diff --git a/tests/regression.at b/tests/regression.at index a4b20de5..9aaedde0 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 @@ -163,7 +162,7 @@ AT_CLEANUP AT_SETUP([Braces parsing]) AT_DATA([input.y], -[[/* Bison used to swallow the character after `}'. */ +[[/* Bison used to swallow the character after '}'. */ %% exp: { tests = {{{{{{{{{{}}}}}}}}}}; }; @@ -185,7 +184,7 @@ AT_CLEANUP AT_SETUP([Duplicate string]) 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 @@ -198,7 +197,7 @@ exp: '(' exp ')' | NUM ; ]]) 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 ]]) AT_CLEANUP @@ -388,14 +387,14 @@ 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: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 %{...%} ]]) @@ -429,7 +428,7 @@ AT_CLEANUP AT_SETUP([Token definitions]) -# Bison managed, when fed with `%token 'f' "f"' to #define 'f'! +# Bison managed, when fed with '%token 'f' "f"' to #define 'f'! AT_DATA_GRAMMAR([input.y], [%{ #include @@ -476,7 +475,7 @@ main (void) # 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:22.8-63: warning: symbol "\\'?\"\a\b\f\n\r\t\v\001\201\001\201??!" used more than once as a literal string ]]) AT_COMPILE([input]) @@ -525,7 +524,7 @@ 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 +532,7 @@ 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 +542,12 @@ 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 +686,7 @@ 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, @@ -826,7 +825,7 @@ m4_define([_AT_DATA_DANCER_Y], [AT_DATA_GRAMMAR([dancer.y], [%{ static int yylex (AT_LALR1_CC_IF([int *], [void])); -AT_LALR1_CC_IF([], +AT_LALR1_CC_IF([#include ], [#include #include static void yyerror (const char *);]) @@ -954,7 +953,7 @@ m4_define([_AT_DATA_EXPECT2_Y], [AT_DATA_GRAMMAR([expect2.y], [%{ static int yylex (AT_LALR1_CC_IF([int *], [void])); -AT_LALR1_CC_IF([], +AT_LALR1_CC_IF([#include ], [#include #include static void yyerror (const char *);]) @@ -1063,7 +1062,7 @@ 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"); } ; @@ -1149,7 +1148,7 @@ 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]], @@ -1197,12 +1196,15 @@ AT_DATA_GRAMMAR([input.y], %} %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" @@ -1232,9 +1234,432 @@ main (void) ]]) 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" +[[input.y:23.5-19: warning: rule useless in parser due to conflicts: start: start +input.y:27.5-19: warning: rule useless in parser due to conflicts: sr_conflict: TK2 "tok alias" ]]) 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 expout] +AT_BISON_CHECK([[-o input.c -Dlr.type=ielr input.y]]) +[mv input.c ielr.c] +AT_CHECK([[cat ielr.c]], [[0]], [[expout]]) + +AT_CLEANUP + + + +## -------------------------------------------- ## +## parse.error=verbose and YYSTACK_USE_ALLOCA. ## +## -------------------------------------------- ## + +AT_SETUP([[parse.error=verbose and YYSTACK_USE_ALLOCA]]) + +AT_DATA_GRAMMAR([input.y], +[[%code { + #include + void yyerror (char const *); + int yylex (void); + #define YYSTACK_USE_ALLOCA 1 +} + +%define parse.error verbose + +%% + +start: check syntax_error syntax_error ; + +check: +{ + 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' +; + +%% + +void +yyerror (char const *msg) +{ + fprintf (stderr, "%s\n", msg); +} + +int +yylex (void) +{ + /* Induce two syntax error messages (which requires full error + recovery by shifting 3 tokens) in order to detect any loss of the + reallocated buffer. */ + static char const *input = "abc"; + return *input++; +} + +int +main (void) +{ + return yyparse (); +} +]]) + +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_DATA_GRAMMAR([input.y], +[[%code { + #include + void yyerror (char const *); + int yylex (void); + + /* 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" +; + +%% + +void +yyerror (char const *msg) +{ + fprintf (stderr, "%s\n", msg); +} + +int +yylex (void) +{ + /* Induce two syntax error messages (which requires full error + recovery by shifting 3 tokens). */ + static char const *input = "abc"; + return *input++; +} + +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]]) + +# 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_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 + void yyerror (char const *); + 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"); } ; + +%% + +void +yyerror (char const *msg) +{ + fprintf (stderr, "%s\n", msg); +} + +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: conflicts: 21 shift/reduce +]]) +AT_COMPILE([[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 + + + +## ------------------------ ## +## LAC: Memory exhaustion. ## +## ------------------------ ## + +AT_SETUP([[LAC: Memory exhaustion]]) + +m4_pushdef([AT_LAC_CHECK], [ + +AT_DATA_GRAMMAR([input.y], +[[%code { + #include + void yyerror (char const *); + int yylex (void); + #define YYMAXDEPTH 8 +} + +%error-verbose + +%% + +S: A A A A A A A A A ; +A: /*empty*/ | 'a' ; + +%% + +void +yyerror (char const *msg) +{ + fprintf (stderr, "%s\n", msg); +} + +int +yylex (void) +{ + static char const *input = "]$1["; + return *input++; +} + +int +main (void) +{ + yydebug = 1; + return yyparse (); +} +]]) + +AT_BISON_CHECK([[-Dparse.lac=full -Dparse.lac.es-capacity-initial=1 \ + -t -o input.c input.y]], [[0]], [], +[[input.y: conflicts: 8 shift/reduce +]]) +AT_COMPILE([[input]]) + +]) + +# 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