Future/bison
1
0
Fork 0
mirror of https://git.savannah.gnu.org/git/bison.git synced 2026-03-09 20:33:03 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
597600fd837906ff4a07e26b9351e8f5e4f575a2
bison/src/gram.h
Paul Hilfinger b34b12c4f9 allow %expect and %expect-rr modifiers on individual rules 
This change allows one to document (and check) which rules participate
in shift/reduce and reduce/reduce conflicts.  This is particularly
important GLR parsers, where conflicts are a normal occurrence.  For
example,

    %glr-parser
    %expect 1
    %%

    ...

    argument_list:
      arguments %expect 1
    | arguments ','
    | %empty
    ;

    arguments:
      expression
    | argument_list ',' expression
    ;

    ...

Looking at the output from -v, one can see that the shift-reduce
conflict here is due to the fact that the parser does not know whether
to reduce arguments to argument_list until it sees the token AFTER the
following ','.  By marking the rule with %expect 1 (because there is a
conflict in one state), we document the source of the 1 overall shift-
reduce conflict.

In GLR parsers, we can use %expect-rr in a rule for reduce/reduce
conflicts.  In this case, we mark each of the conflicting rules.  For
example,

    %glr-parser
    %expect-rr 1

    %%

    stmt:
      target_list '=' expr ';'
    | expr_list ';'
    ;

    target_list:
      target
    | target ',' target_list
    ;

    target:
      ID %expect-rr 1
    ;

    expr_list:
      expr
    | expr ',' expr_list
    ;

    expr:
      ID %expect-rr 1
    | ...
    ;

In a statement such as

    x, y = 3, 4;

the parser must reduce x to a target or an expr, but does not know
which until it sees the '='.  So we notate the two possible reductions
to indicate that each conflicts in one rule.

See https://lists.gnu.org/archive/html/bison-patches/2013-02/msg00105.html.

* doc/bison.texi (Suppressing Conflict Warnings): Document %expect,
%expect-rr in grammar rules.
* src/conflicts.c (count_state_rr_conflicts): Adjust comment.
(rule_has_state_sr_conflicts): New static function.
(count_rule_sr_conflicts): New static function.
(rule_nast_state_rr_conflicts): New static function.
(count_rule_rr_conflicts): New static function.
(rule_conflicts_print): New static function.
(conflicts_print): Also use rule_conflicts_print to report on individual
rules.
* src/gram.h (struct rule): Add new fields expected_sr_conflicts,
expected_rr_conflicts.
* src/reader.c (grammar_midrule_action): Transfer expected_sr_conflicts,
expected_rr_conflicts to new rule, and turn off in current_rule.
(grammar_current_rule_expect_sr): New function.
(grammar_current_rule_expect_rr): New function.
(packgram): Transfer expected_sr_conflicts, expected_rr_conflicts
to new rule.
* src/reader.h (grammar_current_rule_expect_sr): New function.
(grammar_current_rule_expect_rr): New function.
* src/symlist.c (symbol_list_sym_new): Initialize expected_sr_conflicts,
expected_rr_conflicts.
* src/symlist.h (struct symbol_list): Add new fields expected_sr_conflicts,
expected_rr_conflicts.
* tests/conflicts.at: Add tests "%expect in grammar rule not enough",
"%expect in grammar rule right.", "%expect in grammar rule too much."
2018-11-21 22:08:47 +01:00

284 lines
8.6 KiB
C
Raw Blame History

/* Data definitions for internal representation of Bison's input.
Copyright (C) 1984, 1986, 1989, 1992, 2001-2007, 2009-2015, 2018 Free
Software Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
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
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>. */
#ifndef GRAM_H_
# define GRAM_H_
/* Representation of the grammar rules:
NTOKENS is the number of tokens, and NVARS is the number of
variables (nonterminals). NSYMS is the total number, ntokens +
nvars.
Each symbol (either token or variable) receives a symbol number.
Numbers 0 to NTOKENS - 1 are for tokens, and NTOKENS to NSYMS - 1
are for variables. Symbol number zero is the end-of-input token.
This token is counted in ntokens. The true number of token values
assigned is NTOKENS reduced by one for each alias declaration.
The rules receive rule numbers 1 to NRULES in the order they are
written. More precisely Bison augments the grammar with the
initial rule, '$accept: START-SYMBOL $end', which is numbered 1,
all the user rules are 2, 3 etc. Each time a rule number is
presented to the user, we subtract 1, so *displayed* rule numbers
are 0, 1, 2...
Internally, we cannot use the number 0 for a rule because for
instance RITEM stores both symbol (the RHS) and rule numbers: the
symbols are shorts >= 0, and rule number are stored negative.
Therefore 0 cannot be used, since it would be both the rule number
0, and the token $end).
Actions are accessed via the rule number.
The rules themselves are described by several arrays: amongst which
RITEM, and RULES.
RULES is an array of rules, whose members are:
RULES[R].lhs -- the symbol of the left hand side of rule R.
RULES[R].rhs -- the index in RITEM of the beginning of the portion
for rule R.
RULES[R].prec -- the symbol providing the precedence level of R.
RULES[R].precsym -- the symbol attached (via %prec) to give its
precedence to R. Of course, if set, it is equal to 'prec', but we
need to distinguish one from the other when reducing: a symbol used
in a %prec is not useless.
RULES[R].assoc -- the associativity of R.
RULES[R].dprec -- the dynamic precedence level of R (for GLR
parsing).
RULES[R].merger -- index of merging function for R (for GLR
parsing).
RULES[R].line -- the line where R was defined.
RULES[R].useful -- true iff the rule is used (i.e., false if thrown
away by reduce).
The right hand side is stored as symbol numbers in a portion of
RITEM.
The length of the portion is one greater than the number of symbols
in the rule's right hand side. The last element in the portion
contains minus R, which identifies it as the end of a portion and
says which rule it is for.
The portions of RITEM come in order of increasing rule number.
NRITEMS is the total length of RITEM. Each element of RITEM is
called an "item" and its index in RITEM is an item number.
Item numbers are used in the finite state machine to represent
places that parsing can get to.
SYMBOLS[I]->prec records the precedence level of each symbol.
Precedence levels are assigned in increasing order starting with 1
so that numerically higher precedence values mean tighter binding
as they ought to. Zero as a symbol or rule's precedence means none
is assigned.
Associativities are recorded similarly in SYMBOLS[I]->assoc. */
# include "location.h"
# include "symtab.h"
# define ISTOKEN(i) ((i) < ntokens)
# define ISVAR(i) ((i) >= ntokens)
extern int nsyms;
extern int ntokens;
extern int nvars;
typedef int item_number;
# define ITEM_NUMBER_MAX INT_MAX
extern item_number *ritem;
extern unsigned nritems;
/* There is weird relationship between OT1H item_number and OTOH
symbol_number and rule_number: we store the latter in
item_number. symbol_number values are stored as-is, while
the negation of (rule_number + 1) is stored.
Therefore, a symbol_number must be a valid item_number, and we
sometimes have to perform the converse transformation. */
static inline item_number
symbol_number_as_item_number (symbol_number sym)
{
return sym;
}
static inline symbol_number
item_number_as_symbol_number (item_number i)
{
return i;
}
static inline bool
item_number_is_symbol_number (item_number i)
{
return i >= 0;
}
/* Rule numbers. */
typedef int rule_number;
# define RULE_NUMBER_MAX INT_MAX
extern rule_number nrules;
static inline item_number
rule_number_as_item_number (rule_number r)
{
return -1 - r;
}
static inline rule_number
item_number_as_rule_number (item_number i)
{
return -1 - i;
}
static inline bool
item_number_is_rule_number (item_number i)
{
return i < 0;
}
/*--------.
| Rules. |
`--------*/
typedef struct
{
/* The number of the rule in the source. It is usually the index in
RULES too, except if there are useless rules. */
rule_number user_number;
/* The index in RULES. Usually the rule number in the source,
except if some rules are useless. */
rule_number number;
sym_content *lhs;
item_number *rhs;
/* This symbol provides both the associativity, and the precedence. */
sym_content *prec;
int dprec;
int merger;
/* This symbol was attached to the rule via %prec. */
sym_content *precsym;
location location;
bool useful;
bool is_predicate;
/* Counts of the numbers of expected conflicts for this rule, or -1 if none
given. */
int expected_sr_conflicts;
int expected_rr_conflicts;
const char *action;
location action_location;
} rule;
extern rule *rules;
/* A function that selects a rule. */
typedef bool (*rule_filter) (rule const *);
/* Return true IFF the rule has a 'number' smaller than NRULES. That is, it is
useful in the grammar. */
bool rule_useful_in_grammar_p (rule const *r);
/* Return true IFF the rule has a 'number' higher than NRULES. That is, it is
useless in the grammar. */
bool rule_useless_in_grammar_p (rule const *r);
/* Return true IFF the rule is not flagged as useful but is useful in the
grammar. In other words, it was discarded because of conflicts. */
bool rule_useless_in_parser_p (rule const *r);
/* Print this rule's number and lhs on OUT. If a PREVIOUS_LHS was
already displayed (by a previous call for another rule), avoid
useless repetitions. */
void rule_lhs_print (rule const *r, sym_content const *previous_lhs,
FILE *out);
void rule_lhs_print_xml (rule const *r, FILE *out, int level);
/* Return the length of the RHS. */
size_t rule_rhs_length (rule const *r);
/* Print this rule's RHS on OUT. */
void rule_rhs_print (rule const *r, FILE *out);
/* Table of the symbols, indexed by the symbol number. */
extern symbol **symbols;
/* TOKEN_TRANSLATION -- a table indexed by a token number as returned
by the user's yylex routine, it yields the internal token number
used by the parser and throughout bison. */
extern symbol_number *token_translations;
extern int max_user_token_number;
/* Dump RITEM for traces. */
void ritem_print (FILE *out);
/* Return the size of the longest rule RHS. */
size_t ritem_longest_rhs (void);
/* Print the grammar's rules that match FILTER on OUT under TITLE. */
void grammar_rules_partial_print (FILE *out, const char *title,
rule_filter filter);
/* Print the grammar's useful rules on OUT. */
void grammar_rules_print (FILE *out);
/* Print all of the grammar's rules with a "usefulness" attribute. */
void grammar_rules_print_xml (FILE *out, int level);
/* Dump the grammar. */
void grammar_dump (FILE *out, const char *title);
/* Report on STDERR the rules that are not flagged USEFUL, using the
MESSAGE (which can be 'rule useless in grammar' when invoked after grammar
reduction, or 'rule useless in parser due to conflicts' after conflicts
were taken into account). */
void grammar_rules_useless_report (const char *message);
/* Free the packed grammar. */
void grammar_free (void);
/* The version %required by the grammar file, as an int (100 * major +
minor). 0 if unspecified. */
extern int required_version;
#endif /* !GRAM_H_ */
Reference in New Issue View Git Blame Copy Permalink