src/state.h

   1 /* Type definitions for nondeterministic finite state machine for Bison.
   2
   3    Copyright (C) 1984, 1989, 2000, 2001, 2002, 2003, 2004 Free
   4    Software 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., 51 Franklin Street, Fifth Floor,
  21    Boston, MA 02110-1301, 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 RITEM 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.  rules is an array of
  67    num rules.  lookahead_tokens is an array of bitsets, one per rule.
  68
  69    Conflict resolution can decide that certain tokens in certain
  70    states should explicitly be errors (for implementing %nonassoc).
  71    For each state, the tokens that are errors for this reason are
  72    recorded in an errs structure, which holds the token numbers.
  73
  74    There is at least one goto transition present in state zero.  It
  75    leads to a next-to-final state whose accessing_symbol is the
  76    grammar's start symbol.  The next-to-final state has one shift to
  77    the final state, whose accessing_symbol is zero (end of input).
  78    The final state has one shift, which goes to the termination state.
  79    The reason for the extra state at the end is to placate the
  80    parser's strategy of making all decisions one token ahead of its
  81    actions.  */
  82
  83 #ifndef STATE_H_
  84 # define STATE_H_
  85
  86 # include <bitset.h>
  87
  88 # include "gram.h"
  89 # include "symtab.h"
  90
  91
  92 /*-------------------.
  93 | Numbering states.  |
  94 `-------------------*/
  95
  96 typedef int state_number;
  97 # define STATE_NUMBER_MAXIMUM INT_MAX
  98
  99 /* Be ready to map a state_number to an int.  */
 100 static inline int
 101 state_number_as_int (state_number s)
 102 {
 103   return s;
 104 }
 105
 106
 107 typedef struct state state;
 108
 109 /*--------------.
 110 | Transitions.  |
 111 `--------------*/
 112
 113 typedef struct
 114 {
 115   int num;
 116   state *states[1];
 117 } transitions;
 118
 119
 120 /* What is the symbol labelling the transition to
 121    TRANSITIONS->states[Num]?  Can be a token (amongst which the error
 122    token), or non terminals in case of gotos.  */
 123
 124 #define TRANSITION_SYMBOL(Transitions, Num) \
 125   (Transitions->states[Num]->accessing_symbol)
 126
 127 /* Is the TRANSITIONS->states[Num] a shift? (as opposed to gotos).  */
 128
 129 #define TRANSITION_IS_SHIFT(Transitions, Num) \
 130   (ISTOKEN (TRANSITION_SYMBOL (Transitions, Num)))
 131
 132 /* Is the TRANSITIONS->states[Num] a goto?. */
 133
 134 #define TRANSITION_IS_GOTO(Transitions, Num) \
 135   (!TRANSITION_IS_SHIFT (Transitions, Num))
 136
 137 /* Is the TRANSITIONS->states[Num] labelled by the error token?  */
 138
 139 #define TRANSITION_IS_ERROR(Transitions, Num) \
 140   (TRANSITION_SYMBOL (Transitions, Num) == errtoken->number)
 141
 142 /* When resolving a SR conflicts, if the reduction wins, the shift is
 143    disabled.  */
 144
 145 #define TRANSITION_DISABLE(Transitions, Num) \
 146   (Transitions->states[Num] = NULL)
 147
 148 #define TRANSITION_IS_DISABLED(Transitions, Num) \
 149   (Transitions->states[Num] == NULL)
 150
 151
 152 /* Iterate over each transition over a token (shifts).  */
 153 #define FOR_EACH_SHIFT(Transitions, Iter)                       \
 154   for (Iter = 0;                                                \
 155        Iter < Transitions->num                                  \
 156          && (TRANSITION_IS_DISABLED (Transitions, Iter)         \
 157              || TRANSITION_IS_SHIFT (Transitions, Iter));       \
 158        ++Iter)                                                  \
 159     if (!TRANSITION_IS_DISABLED (Transitions, Iter))
 160
 161
 162 /* Return the state such SHIFTS contain a shift/goto to it on SYM.
 163    Abort if none found.  */
 164 struct state *transitions_to (transitions *shifts, symbol_number sym);
 165
 166
 167 /*-------.
 168 | Errs.  |
 169 `-------*/
 170
 171 typedef struct
 172 {
 173   int num;
 174   symbol *symbols[1];
 175 } errs;
 176
 177 errs *errs_new (int num, symbol **tokens);
 178
 179
 180 /*-------------.
 181 | Reductions.  |
 182 `-------------*/
 183
 184 typedef struct
 185 {
 186   int num;
 187   bitset *lookahead_tokens;
 188   rule *rules[1];
 189 } reductions;
 190
 191
 192
 193 /*---------.
 194 | states.  |
 195 `---------*/
 196
 197 struct state
 198 {
 199   state_number number;
 200   symbol_number accessing_symbol;
 201   transitions *transitions;
 202   reductions *reductions;
 203   errs *errs;
 204
 205   /* Nonzero if no lookahead is needed to decide what to do in state S.  */
 206   char consistent;
 207
 208   /* If some conflicts were solved thanks to precedence/associativity,
 209      a human readable description of the resolution.  */
 210   const char *solved_conflicts;
 211
 212   /* Conflict resolution sometimes makes states unreachable.  Initialized to 0
 213      in state_new and then used by state_remove_unreachable_states after
 214      conflicts_solve.  */
 215   bool reachable;
 216
 217   /* Its items.  Must be last, since ITEMS can be arbitrarily large.
 218      */
 219   size_t nitems;
 220   item_number items[1];
 221 };
 222
 223 extern state_number nstates;
 224 extern state *final_state;
 225
 226 /* Create a new state with ACCESSING_SYMBOL for those items.  */
 227 state *state_new (symbol_number accessing_symbol,
 228                   size_t core_size, item_number *core);
 229
 230 /* Set the transitions of STATE.  */
 231 void state_transitions_set (state *s, int num, state **trans);
 232
 233 /* Set the reductions of STATE.  */
 234 void state_reductions_set (state *s, int num, rule **reds);
 235
 236 int state_reduction_find (state *s, rule *r);
 237
 238 /* Set the errs of STATE.  */
 239 void state_errs_set (state *s, int num, symbol **errors);
 240
 241 /* Print on OUT all the lookahead tokens such that this STATE wants to
 242    reduce R.  */
 243 void state_rule_lookahead_tokens_print (state *s, rule *r, FILE *out);
 244
 245 /* Create/destroy the states hash table.  */
 246 void state_hash_new (void);
 247 void state_hash_free (void);
 248
 249 /* Find the state associated to the CORE, and return it.  If it does
 250    not exist yet, return NULL.  */
 251 state *state_hash_lookup (size_t core_size, item_number *core);
 252
 253 /* Insert STATE in the state hash table.  */
 254 void state_hash_insert (state *s);
 255
 256 /* Remove unreachable states, renumber remaining states, update NSTATES, and
 257    write to OLD_TO_NEW a mapping of old state numbers to new state numbers such
 258    that the old value of NSTATES is written as the new state number for removed
 259    states.  The size of OLD_TO_NEW must be the old value of NSTATES.  */
 260 void state_remove_unreachable_states (state_number old_to_new[]);
 261
 262 /* All the states, indexed by the state number.  */
 263 extern state **states;
 264
 265 /* Free all the states.  */
 266 void states_free (void);
 267
 268 #endif /* !STATE_H_ */