src/state.h

   1 /* Type definitions for nondeterministic finite state machine for bison,
   2
   3    Copyright (C) 1984, 1989, 2000, 2001, 2002 Free Software
   4    Foundation, Inc.
   5
   6    This file is part of Bison, the GNU Compiler Compiler.
   7
   8    Bison is free software; you can redistribute it and/or modify
   9    it under the terms of the GNU General Public License as published by
  10    the Free Software Foundation; either version 2, or (at your option)
  11    any later version.
  12
  13    Bison is distributed in the hope that it will be useful,
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16    GNU General Public License for more details.
  17
  18    You should have received a copy of the GNU General Public License
  19    along with Bison; see the file COPYING.  If not, write to
  20    the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  21    Boston, MA 02111-1307, USA.  */
  22
  23
  24 /* These type definitions are used to represent a nondeterministic
  25    finite state machine that parses the specified grammar.  This
  26    information is generated by the function generate_states in the
  27    file LR0.
  28
  29    Each state of the machine is described by a set of items --
  30    particular positions in particular rules -- that are the possible
  31    places where parsing could continue when the machine is in this
  32    state.  These symbols at these items are the allowable inputs that
  33    can follow now.
  34
  35    A core represents one state.  States are numbered in the NUMBER
  36    field.  When generate_states is finished, the starting state is
  37    state 0 and NSTATES is the number of states.  (FIXME: This sentence
  38    is no longer true: A transition to a state whose state number is
  39    NSTATES indicates termination.)  All the cores are chained together
  40    and FIRST_STATE points to the first one (state 0).
  41
  42    For each state there is a particular symbol which must have been
  43    the last thing accepted to reach that state.  It is the
  44    ACCESSING_SYMBOL of the core.
  45
  46    Each core contains a vector of NITEMS items which are the indices
  47    in the RITEMS vector of the items that are selected in this state.
  48
  49    The two types of actions are shifts/gotos (push the lookahead token
  50    and read another/goto to the state designated by a nterm) and
  51    reductions (combine the last n things on the stack via a rule,
  52    replace them with the symbol that the rule derives, and leave the
  53    lookahead token alone).  When the states are generated, these
  54    actions are represented in two other lists.
  55
  56    Each transition structure describes the possible transitions out
  57    of one state, the state whose number is in the number field.  Each
  58    contains a vector of numbers of the states that transitions can go
  59    to.  The accessing_symbol fields of those states' cores say what
  60    kind of input leads to them.
  61
  62    A transition to state zero should be ignored: conflict resolution
  63    deletes transitions by having them point to zero.
  64
  65    Each reductions structure describes the possible reductions at the
  66    state whose number is in the number field.  The data is a list of
  67    nreds rules, represented by their rule numbers.  first_reduction
  68    points to the list of these structures.
  69
  70    Conflict resolution can decide that certain tokens in certain
  71    states should explicitly be errors (for implementing %nonassoc).
  72    For each state, the tokens that are errors for this reason are
  73    recorded in an errs structure, which holds the token numbers.
  74
  75    There is at least one goto transition present in state zero.  It
  76    leads to a next-to-final state whose accessing_symbol is the
  77    grammar's start symbol.  The next-to-final state has one shift to
  78    the final state, whose accessing_symbol is zero (end of input).
  79    The final state has one shift, which goes to the termination state.
  80    The reason for the extra state at the end is to placate the
  81    parser's strategy of making all decisions one token ahead of its
  82    actions.  */
  83
  84 #ifndef STATE_H_
  85 # define STATE_H_
  86
  87 # include <bitset.h>
  88
  89 # include "gram.h"
  90 # include "symtab.h"
  91
  92
  93 /*-------------------.
  94 | Numbering states.  |
  95 `-------------------*/
  96
  97 typedef short state_number;
  98 # define STATE_NUMBER_MAXIMUM SHRT_MAX
  99
 100 /* Be ready to map a state_number to an int.  */
 101 # define state_number_as_int(Tok) ((int) (Tok))
 102
 103
 104 typedef struct state state;
 105
 106 /*--------------.
 107 | Transitions.  |
 108 `--------------*/
 109
 110 typedef struct
 111 {
 112   short num;
 113   state *states[1];
 114 } transitions;
 115
 116
 117 /* What is the symbol labelling the transition to
 118    TRANSITIONS->states[Num]?  Can be a token (amongst which the error
 119    token), or non terminals in case of gotos.  */
 120
 121 #define TRANSITION_SYMBOL(Transitions, Num) \
 122   (Transitions->states[Num]->accessing_symbol)
 123
 124 /* Is the TRANSITIONS->states[Num] a shift? (as opposed to gotos).  */
 125
 126 #define TRANSITION_IS_SHIFT(Transitions, Num) \
 127   (ISTOKEN (TRANSITION_SYMBOL (Transitions, Num)))
 128
 129 /* Is the TRANSITIONS->states[Num] a goto?. */
 130
 131 #define TRANSITION_IS_GOTO(Transitions, Num) \
 132   (!TRANSITION_IS_SHIFT (Transitions, Num))
 133
 134 /* Is the TRANSITIONS->states[Num] labelled by the error token?  */
 135
 136 #define TRANSITION_IS_ERROR(Transitions, Num) \
 137   (TRANSITION_SYMBOL (Transitions, Num) == errtoken->number)
 138
 139 /* When resolving a SR conflicts, if the reduction wins, the shift is
 140    disabled.  */
 141
 142 #define TRANSITION_DISABLE(Transitions, Num) \
 143   (Transitions->states[Num] = NULL)
 144
 145 #define TRANSITION_IS_DISABLED(Transitions, Num) \
 146   (Transitions->states[Num] == NULL)
 147
 148
 149 /* Iterate over each transition over a token (shifts).  */
 150 #define FOR_EACH_SHIFT(Transitions, Iter)                       \
 151   for (Iter = 0;                                                \
 152        Iter < Transitions->num                                  \
 153          && (TRANSITION_IS_DISABLED (Transitions, Iter)         \
 154              || TRANSITION_IS_SHIFT (Transitions, Iter));       \
 155        ++Iter)                                                  \
 156     if (!TRANSITION_IS_DISABLED (Transitions, Iter))
 157
 158
 159 /* Return the state such SHIFTS contain a shift/goto to it on S.
 160    Abort if none found.  */
 161 struct state *transitions_to (transitions *shifts, symbol_number s);
 162
 163
 164 /*-------.
 165 | Errs.  |
 166 `-------*/
 167
 168 typedef struct
 169 {
 170   short num;
 171   symbol *symbols[1];
 172 } errs;
 173
 174 errs *errs_new (int num, symbol **tokens);
 175
 176
 177 /*-------------.
 178 | Reductions.  |
 179 `-------------*/
 180
 181 typedef struct
 182 {
 183   short num;
 184   bitset *lookaheads;
 185   rule *rules[1];
 186 } reductions;
 187
 188
 189
 190 /*---------.
 191 | states.  |
 192 `---------*/
 193
 194 struct state
 195 {
 196   state_number number;
 197   symbol_number accessing_symbol;
 198   transitions *transitions;
 199   reductions *reductions;
 200   errs *errs;
 201
 202   /* Nonzero if no lookahead is needed to decide what to do in state S.  */
 203   char consistent;
 204
 205   /* If some conflicts were solved thanks to precedence/associativity,
 206      a human readable description of the resolution.  */
 207   const char *solved_conflicts;
 208
 209   /* Its items.  Must be last, since ITEMS can be arbitrarily large.
 210      */
 211   unsigned short nitems;
 212   item_number items[1];
 213 };
 214
 215 extern state_number nstates;
 216 extern state *final_state;
 217
 218 /* Create a new state with ACCESSING_SYMBOL for those items.  */
 219 state *state_new (symbol_number accessing_symbol,
 220                   size_t core_size, item_number *core);
 221
 222 /* Set the transitions of STATE.  */
 223 void state_transitions_set (state *s, int num, state **trans);
 224
 225 /* Set the reductions of STATE.  */
 226 void state_reductions_set (state *s, int num, rule **reds);
 227
 228 int state_reduction_find (state *s, rule *r);
 229
 230 /* Set the errs of STATE.  */
 231 void state_errs_set (state *s, int num, symbol **errors);
 232
 233 /* Print on OUT all the lookaheads such that this STATE wants to
 234    reduce R.  */
 235 void state_rule_lookaheads_print (state *s, rule *r, FILE *out);
 236
 237 /* Create/destroy the states hash table.  */
 238 void state_hash_new (void);
 239 void state_hash_free (void);
 240
 241 /* Find the state associated to the CORE, and return it.  If it does
 242    not exist yet, return NULL.  */
 243 state *state_hash_lookup (size_t core_size, item_number *core);
 244
 245 /* Insert STATE in the state hash table.  */
 246 void state_hash_insert (state *s);
 247
 248 /* All the states, indexed by the state number.  */
 249 extern state **states;
 250
 251 /* Free all the states.  */
 252 void states_free (void);
 253 #endif /* !STATE_H_ */