/* Generate the nondeterministic finite state machine for bison,
- Copyright 1984, 1986, 1989, 2000 Free Software Foundation, Inc.
+ Copyright 1984, 1986, 1989, 2000, 2001, 2002 Free Software Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
The entry point is generate_states. */
#include "system.h"
+#include "bitset.h"
+#include "quotearg.h"
+#include "symtab.h"
+#include "gram.h"
+#include "getargs.h"
+#include "reader.h"
#include "gram.h"
#include "state.h"
#include "complain.h"
#include "closure.h"
#include "LR0.h"
+#include "lalr.h"
+#include "reduce.h"
+unsigned int nstates = 0;
+/* Initialize the final state to -1, otherwise, it might be set to 0
+ by default, and since we don't compute the reductions of the final
+ state, we end up not computing the reductions of the initial state,
+ which is of course needed.
-int nstates;
-int final_state;
-core *first_state = NULL;
-shifts *first_shift = NULL;
-reductions *first_reduction = NULL;
+ FINAL_STATE is properly set by new_state when it recognizes the
+ accessing symbol: EOF. */
+int final_state = -1;
+static state_t *first_state = NULL;
-static core *this_state = NULL;
-static core *last_state = NULL;
-static shifts *last_shift = NULL;
-static reductions *last_reduction = NULL;
+static state_t *this_state = NULL;
+static state_t *last_state = NULL;
static int nshifts;
-static short *shift_symbol = NULL;
+static token_number_t *shift_symbol = NULL;
static short *redset = NULL;
static short *shiftset = NULL;
-static short **kernel_base = NULL;
-static short **kernel_end = NULL;
-static short *kernel_items = NULL;
+static item_number_t **kernel_base = NULL;
+static int *kernel_size = NULL;
+static item_number_t *kernel_items = NULL;
/* hash table for states, to recognize equivalent ones. */
-#define STATE_TABLE_SIZE 1009
-static core **state_table = NULL;
+#define STATE_HASH_SIZE 1009
+static state_t **state_hash = NULL;
\f
static void
allocate_itemsets (void)
{
- short *itemp = NULL;
- int symbol;
- int i;
- int count;
- short *symbol_count = NULL;
-
- count = 0;
- symbol_count = XCALLOC (short, nsyms);
-
- itemp = ritem;
- symbol = *itemp++;
- while (symbol)
- {
- if (symbol > 0)
- {
- count++;
- symbol_count[symbol]++;
- }
- symbol = *itemp++;
- }
+ int i, r;
+ item_number_t *rhsp;
+
+ /* Count the number of occurrences of all the symbols in RITEMS.
+ Note that useless productions (hence useless nonterminals) are
+ browsed too, hence we need to allocate room for _all_ the
+ symbols. */
+ int count = 0;
+ short *symbol_count = XCALLOC (short, nsyms + nuseless_nonterminals);
+
+ for (r = 1; r < nrules + 1; ++r)
+ for (rhsp = rules[r].rhs; *rhsp >= 0; ++rhsp)
+ {
+ count++;
+ symbol_count[*rhsp]++;
+ }
/* See comments before new_itemsets. All the vectors of items
live inside KERNEL_ITEMS. The number of active items after
appears as an item, which is symbol_count[symbol].
We allocate that much space for each symbol. */
- kernel_base = XCALLOC (short *, nsyms);
+ kernel_base = XCALLOC (item_number_t *, nsyms);
if (count)
- kernel_items = XCALLOC (short, count);
+ kernel_items = XCALLOC (item_number_t, count);
count = 0;
for (i = 0; i < nsyms; i++)
count += symbol_count[i];
}
- shift_symbol = symbol_count;
- kernel_end = XCALLOC (short *, nsyms);
+ free (symbol_count);
+ kernel_size = XCALLOC (int, nsyms);
}
shiftset = XCALLOC (short, nsyms);
redset = XCALLOC (short, nrules + 1);
- state_table = XCALLOC (core *, STATE_TABLE_SIZE);
+ state_hash = XCALLOC (state_t *, STATE_HASH_SIZE);
+ shift_symbol = XCALLOC (token_number_t, nsyms);
}
static void
free_storage (void)
{
- XFREE (shift_symbol);
- XFREE (redset);
- XFREE (shiftset);
- XFREE (kernel_base);
- XFREE (kernel_end);
+ free (shift_symbol);
+ free (redset);
+ free (shiftset);
+ free (kernel_base);
+ free (kernel_size);
XFREE (kernel_items);
- XFREE (state_table);
+ free (state_hash);
}
| shift_symbol is set to a vector of the symbols that can be |
| shifted. For each symbol in the grammar, kernel_base[symbol] |
| points to a vector of item numbers activated if that symbol is |
-| shifted, and kernel_end[symbol] points after the end of that |
-| vector. |
+| shifted, and kernel_size[symbol] is their numbers. |
`----------------------------------------------------------------*/
static void
new_itemsets (void)
{
int i;
- int shiftcount;
- short *isp;
- short *ksp;
- int symbol;
-#if TRACE
- fprintf (stderr, "Entering new_itemsets\n");
-#endif
+ if (trace_flag)
+ fprintf (stderr, "Entering new_itemsets, state = %d\n",
+ this_state->number);
for (i = 0; i < nsyms; i++)
- kernel_end[i] = NULL;
-
- shiftcount = 0;
-
- isp = itemset;
-
- while (isp < itemsetend)
- {
- i = *isp++;
- symbol = ritem[i];
- if (symbol > 0)
- {
- ksp = kernel_end[symbol];
-
- if (!ksp)
- {
- shift_symbol[shiftcount++] = symbol;
- ksp = kernel_base[symbol];
- }
-
- *ksp++ = i + 1;
- kernel_end[symbol] = ksp;
- }
- }
-
- nshifts = shiftcount;
+ kernel_size[i] = 0;
+
+ nshifts = 0;
+
+ for (i = 0; i < nritemset; ++i)
+ if (ritem[itemset[i]] >= 0)
+ {
+ token_number_t symbol
+ = item_number_as_token_number (ritem[itemset[i]]);
+ if (!kernel_size[symbol])
+ {
+ shift_symbol[nshifts] = symbol;
+ nshifts++;
+ }
+
+ kernel_base[symbol][kernel_size[symbol]] = itemset[i] + 1;
+ kernel_size[symbol]++;
+ }
}
| necessary. |
`-----------------------------------------------------------------*/
-static core *
-new_state (int symbol)
+static state_t *
+new_state (token_number_t symbol)
{
- int n;
- core *p;
- short *isp1;
- short *isp2;
- short *iend;
+ state_t *p;
-#if TRACE
- fprintf (stderr, "Entering new_state, symbol = %d\n", symbol);
-#endif
+ if (trace_flag)
+ fprintf (stderr, "Entering new_state, state = %d, symbol = %d (%s)\n",
+ nstates, symbol, quotearg_style (escape_quoting_style,
+ symbols[symbol]->tag));
- if (nstates >= MAXSHORT)
- fatal (_("too many states (max %d)"), MAXSHORT);
+ if (nstates >= SHRT_MAX)
+ fatal (_("too many states (max %d)"), SHRT_MAX);
- isp1 = kernel_base[symbol];
- iend = kernel_end[symbol];
- n = iend - isp1;
-
- p =
- (core *) xcalloc ((unsigned) (sizeof (core) + (n - 1) * sizeof (short)), 1);
+ p = STATE_ALLOC (kernel_size[symbol]);
p->accessing_symbol = symbol;
p->number = nstates;
- p->nitems = n;
+ p->nitems = kernel_size[symbol];
+
+ memcpy (p->items, kernel_base[symbol],
+ kernel_size[symbol] * sizeof (kernel_base[symbol][0]));
- isp2 = p->items;
- while (isp1 < iend)
- *isp2++ = *isp1++;
+ /* If this is the eoftoken, and this is not the initial state, then
+ this is the final state. */
+ if (symbol == 0 && first_state)
+ final_state = p->number;
- last_state->next = p;
+ if (!first_state)
+ first_state = p;
+ if (last_state)
+ last_state->next = p;
last_state = p;
nstates++;
/*--------------------------------------------------------------.
| Find the state number for the state we would get to (from the |
| current state) by shifting symbol. Create a new state if no |
-| equivalent one exists already. Used by append_states. |
+| equivalent one exists already. Used by append_states. |
`--------------------------------------------------------------*/
static int
-get_state (int symbol)
+get_state (token_number_t symbol)
{
int key;
- short *isp1;
- short *isp2;
- short *iend;
- core *sp;
- int found;
-
- int n;
-
-#if TRACE
- fprintf (stderr, "Entering get_state, symbol = %d\n", symbol);
-#endif
+ int i;
+ state_t *sp;
- isp1 = kernel_base[symbol];
- iend = kernel_end[symbol];
- n = iend - isp1;
+ if (trace_flag)
+ fprintf (stderr, "Entering get_state, state = %d, symbol = %d (%s)\n",
+ this_state->number, symbol, quotearg_style (escape_quoting_style,
+ symbols[symbol]->tag));
- /* add up the target state's active item numbers to get a hash key */
+ /* Add up the target state's active item numbers to get a hash key.
+ */
key = 0;
- while (isp1 < iend)
- key += *isp1++;
-
- key = key % STATE_TABLE_SIZE;
-
- sp = state_table[key];
+ for (i = 0; i < kernel_size[symbol]; ++i)
+ key += kernel_base[symbol][i];
+ key = key % STATE_HASH_SIZE;
+ sp = state_hash[key];
if (sp)
{
- found = 0;
+ int found = 0;
while (!found)
{
- if (sp->nitems == n)
+ if (sp->nitems == kernel_size[symbol])
{
found = 1;
- isp1 = kernel_base[symbol];
- isp2 = sp->items;
-
- while (found && isp1 < iend)
- {
- if (*isp1++ != *isp2++)
- found = 0;
- }
+ for (i = 0; i < kernel_size[symbol]; ++i)
+ if (kernel_base[symbol][i] != sp->items[i])
+ found = 0;
}
if (!found)
}
else /* bucket is empty */
{
- state_table[key] = sp = new_state (symbol);
+ state_hash[key] = sp = new_state (symbol);
}
+ if (trace_flag)
+ fprintf (stderr, "Exiting get_state => %d\n", sp->number);
+
return sp->number;
}
{
int i;
int j;
- int symbol;
+ token_number_t symbol;
-#if TRACE
- fprintf (stderr, "Entering append_states\n");
-#endif
+ if (trace_flag)
+ fprintf (stderr, "Entering append_states, state = %d\n",
+ this_state->number);
/* first sort shift_symbol into increasing order */
}
for (i = 0; i < nshifts; i++)
- {
- symbol = shift_symbol[i];
- shiftset[i] = get_state (symbol);
- }
+ shiftset[i] = get_state (shift_symbol[i]);
}
static void
new_states (void)
{
- core *p;
-
- p = (core *) xcalloc ((unsigned) (sizeof (core) - sizeof (short)), 1);
- first_state = last_state = this_state = p;
- nstates = 1;
-}
-
-
-static void
-save_shifts (void)
-{
- shifts *p;
- short *sp1;
- short *sp2;
- short *send;
-
- p = (shifts *) xcalloc ((unsigned) (sizeof (shifts) +
- (nshifts - 1) * sizeof (short)), 1);
-
- p->number = this_state->number;
- p->nshifts = nshifts;
-
- sp1 = shiftset;
- sp2 = p->shifts;
- send = shiftset + nshifts;
-
- while (sp1 < send)
- *sp2++ = *sp1++;
-
- if (last_shift)
- {
- last_shift->next = p;
- last_shift = p;
- }
- else
- {
- first_shift = p;
- last_shift = p;
- }
-}
-
-
-/*------------------------------------------------------------------.
-| Subroutine of augment_automaton. Create the next-to-final state, |
-| to which a shift has already been made in the initial state. |
-`------------------------------------------------------------------*/
-
-static void
-insert_start_shift (void)
-{
- core *statep;
- shifts *sp;
-
- statep = (core *) xcalloc ((unsigned) (sizeof (core) - sizeof (short)), 1);
- statep->number = nstates;
- statep->accessing_symbol = start_symbol;
-
- last_state->next = statep;
- last_state = statep;
-
- /* Make a shift from this state to (what will be) the final state. */
- sp = XCALLOC (shifts, 1);
- sp->number = nstates++;
- sp->nshifts = 1;
- sp->shifts[0] = nstates;
-
- last_shift->next = sp;
- last_shift = sp;
+ /* The 0 at the lhs is the index of the item of this initial rule. */
+ kernel_base[0][0] = 0;
+ kernel_size[0] = 1;
+ this_state = new_state (0);
}
-/*------------------------------------------------------------------.
-| Make sure that the initial state has a shift that accepts the |
-| grammar's start symbol and goes to the next-to-final state, which |
-| has a shift going to the final state, which has a shift to the |
-| termination state. Create such states and shifts if they don't |
-| happen to exist already. |
-`------------------------------------------------------------------*/
+/*------------------------------------------------------------.
+| Save the NSHIFTS of SHIFTSET into the current linked list. |
+`------------------------------------------------------------*/
static void
-augment_automaton (void)
+save_shifts (void)
{
- int i;
- int k;
- core *statep;
- shifts *sp;
- shifts *sp2;
- shifts *sp1 = NULL;
-
- sp = first_shift;
-
- if (sp)
- {
- if (sp->number == 0)
- {
- k = sp->nshifts;
- statep = first_state->next;
-
- /* The states reached by shifts from first_state are numbered 1...K.
- Look for one reached by start_symbol. */
- while (statep->accessing_symbol < start_symbol
- && statep->number < k)
- statep = statep->next;
-
- if (statep->accessing_symbol == start_symbol)
- {
- /* We already have a next-to-final state.
- Make sure it has a shift to what will be the final state. */
- k = statep->number;
-
- while (sp && sp->number < k)
- {
- sp1 = sp;
- sp = sp->next;
- }
-
- if (sp && sp->number == k)
- {
- sp2 = (shifts *) xcalloc ((unsigned) (sizeof (shifts)
- +
- sp->nshifts *
- sizeof (short)), 1);
- sp2->number = k;
- sp2->nshifts = sp->nshifts + 1;
- sp2->shifts[0] = nstates;
- for (i = sp->nshifts; i > 0; i--)
- sp2->shifts[i] = sp->shifts[i - 1];
-
- /* Patch sp2 into the chain of shifts in place of sp,
- following sp1. */
- sp2->next = sp->next;
- sp1->next = sp2;
- if (sp == last_shift)
- last_shift = sp2;
- XFREE (sp);
- }
- else
- {
- sp2 = XCALLOC (shifts, 1);
- sp2->number = k;
- sp2->nshifts = 1;
- sp2->shifts[0] = nstates;
-
- /* Patch sp2 into the chain of shifts between sp1 and sp. */
- sp2->next = sp;
- sp1->next = sp2;
- if (sp == 0)
- last_shift = sp2;
- }
- }
- else
- {
- /* There is no next-to-final state as yet. */
- /* Add one more shift in first_shift,
- going to the next-to-final state (yet to be made). */
- sp = first_shift;
-
- sp2 = (shifts *) xcalloc (sizeof (shifts)
- + sp->nshifts * sizeof (short), 1);
- sp2->nshifts = sp->nshifts + 1;
-
- /* Stick this shift into the vector at the proper place. */
- statep = first_state->next;
- for (k = 0, i = 0; i < sp->nshifts; k++, i++)
- {
- if (statep->accessing_symbol > start_symbol && i == k)
- sp2->shifts[k++] = nstates;
- sp2->shifts[k] = sp->shifts[i];
- statep = statep->next;
- }
- if (i == k)
- sp2->shifts[k++] = nstates;
-
- /* Patch sp2 into the chain of shifts
- in place of sp, at the beginning. */
- sp2->next = sp->next;
- first_shift = sp2;
- if (last_shift == sp)
- last_shift = sp2;
-
- XFREE (sp);
-
- /* Create the next-to-final state, with shift to
- what will be the final state. */
- insert_start_shift ();
- }
- }
- else
- {
- /* The initial state didn't even have any shifts.
- Give it one shift, to the next-to-final state. */
- sp = XCALLOC (shifts, 1);
- sp->nshifts = 1;
- sp->shifts[0] = nstates;
-
- /* Patch sp into the chain of shifts at the beginning. */
- sp->next = first_shift;
- first_shift = sp;
-
- /* Create the next-to-final state, with shift to
- what will be the final state. */
- insert_start_shift ();
- }
- }
- else
- {
- /* There are no shifts for any state.
- Make one shift, from the initial state to the next-to-final state. */
-
- sp = XCALLOC (shifts, 1);
- sp->nshifts = 1;
- sp->shifts[0] = nstates;
-
- /* Initialize the chain of shifts with sp. */
- first_shift = sp;
- last_shift = sp;
-
- /* Create the next-to-final state, with shift to
- what will be the final state. */
- insert_start_shift ();
- }
-
- /* Make the final state--the one that follows a shift from the
- next-to-final state.
- The symbol for that shift is 0 (end-of-file). */
- statep = (core *) xcalloc ((unsigned) (sizeof (core) - sizeof (short)), 1);
- statep->number = nstates;
- last_state->next = statep;
- last_state = statep;
-
- /* Make the shift from the final state to the termination state. */
- sp = XCALLOC (shifts, 1);
- sp->number = nstates++;
- sp->nshifts = 1;
- sp->shifts[0] = nstates;
- last_shift->next = sp;
- last_shift = sp;
-
- /* Note that the variable `final_state' refers to what we sometimes call
- the termination state. */
- final_state = nstates;
-
- /* Make the termination state. */
- statep = (core *) xcalloc ((unsigned) (sizeof (core) - sizeof (short)), 1);
- statep->number = nstates++;
- last_state->next = statep;
- last_state = statep;
+ shifts *p = shifts_new (nshifts);
+ memcpy (p->shifts, shiftset, nshifts * sizeof (shiftset[0]));
+ this_state->shifts = p;
}
static void
save_reductions (void)
{
- short *isp;
- short *rp1;
- short *rp2;
- int item;
- int count;
- reductions *p;
+ int count = 0;
+ int i;
- short *rend;
+ /* If this is the final state, we want it to have no reductions at
+ all, although it has one for `START_SYMBOL EOF .'. */
+ if (this_state->number == final_state)
+ return;
/* Find and count the active items that represent ends of rules. */
-
- count = 0;
- for (isp = itemset; isp < itemsetend; isp++)
+ for (i = 0; i < nritemset; ++i)
{
- item = ritem[*isp];
+ int item = ritem[itemset[i]];
if (item < 0)
redset[count++] = -item;
}
/* Make a reductions structure and copy the data into it. */
+ this_state->reductions = reductions_new (count);
+ memcpy (this_state->reductions->rules, redset, count * sizeof (redset[0]));
+}
- if (count)
- {
- p = (reductions *) xcalloc ((unsigned) (sizeof (reductions) +
- (count - 1) * sizeof (short)), 1);
-
- p->number = this_state->number;
- p->nreds = count;
-
- rp1 = redset;
- rp2 = p->rules;
- rend = rp1 + count;
+\f
+/*---------------.
+| Build STATES. |
+`---------------*/
- for (/* nothing */; rp1 < rend; ++rp1, ++rp2)
- *rp2 = *rp1;
+static void
+set_states (void)
+{
+ state_t *sp;
+ states = XCALLOC (state_t *, nstates);
- if (last_reduction)
- {
- last_reduction->next = p;
- last_reduction = p;
- }
- else
- {
- first_reduction = p;
- last_reduction = p;
- }
+ for (sp = first_state; sp; sp = sp->next)
+ {
+ /* Pessimization, but simplification of the code: make sure all
+ the states have a shifts, errs, and reductions, even if
+ reduced to 0. */
+ if (!sp->shifts)
+ sp->shifts = shifts_new (0);
+ if (!sp->errs)
+ sp->errs = errs_new (0);
+ if (!sp->reductions)
+ sp->reductions = reductions_new (0);
+
+ states[sp->number] = sp;
}
}
-\f
/*-------------------------------------------------------------------.
| Compute the nondeterministic finite state machine (see state.h for |
| details) from the grammar. |
generate_states (void)
{
allocate_storage ();
- new_closure (nitems);
+ new_closure (nritems);
new_states ();
while (this_state)
{
+ if (trace_flag)
+ fprintf (stderr, "Processing state %d (reached by %s)\n",
+ this_state->number,
+ quotearg_style (escape_quoting_style,
+ symbols[this_state->accessing_symbol]->tag));
/* Set up ruleset and itemset for the transitions out of this
state. ruleset gets a 1 bit for each rule that could reduce
now. itemset gets a vector of all the items that could be
/* create the shifts structures for the shifts to those states,
now that the state numbers transitioning to are known */
- if (nshifts > 0)
- save_shifts ();
+ save_shifts ();
/* states are queued when they are created; process them all */
this_state = this_state->next;
free_closure ();
free_storage ();
- /* set up initial and final states as parser wants them */
- augment_automaton ();
+ /* Set up STATES. */
+ set_states ();
}
projects / bison.git / blobdiff