-/* Compute look-ahead criteria for bison,
- Copyright 1984, 1986, 1989, 2000, 2001 Free Software Foundation, Inc.
+/* Compute look-ahead criteria for Bison.
+
+ Copyright (C) 1984, 1986, 1989, 2000, 2001, 2002, 2003, 2004
+ Free Software Foundation, Inc.
This file is part of Bison, the GNU Compiler Compiler.
/* Compute how to make the finite state machine deterministic; find
- which rules need lookahead in each state, and which lookahead
+ which rules need look-ahead in each state, and which look-ahead
tokens they accept. */
#include "system.h"
-#include "types.h"
+
+#include <bitset.h>
+#include <bitsetv.h>
+#include <quotearg.h>
+
#include "LR0.h"
-#include "gram.h"
#include "complain.h"
-#include "lalr.h"
-#include "nullable.h"
#include "derives.h"
#include "getargs.h"
+#include "gram.h"
+#include "lalr.h"
+#include "nullable.h"
+#include "reader.h"
+#include "relation.h"
+#include "symtab.h"
-/* All the decorated states, indexed by the state number. Warning:
- there is a state_TABLE in LR0.c, but it is different and static.
- */
-state_t **state_table = NULL;
-
-int tokensetsize;
-short *LAruleno;
-unsigned *LA;
-
-static int ngotos;
-short *goto_map;
-short *from_state;
-short *to_state;
-
-/* And for the famous F variable, which name is so descriptive that a
- comment is hardly needed. <grin>. */
-static unsigned *F = NULL;
-#define F(Rule) (F + (Rule) * tokensetsize)
-
-static short **includes;
-static shorts **lookback;
-
-
-/*---------------------------------------------------------------.
-| digraph & traverse. |
-| |
-| The following variables are used as common storage between the |
-| two. |
-`---------------------------------------------------------------*/
-
-static short **R;
-static short *INDEX;
-static short *VERTICES;
-static int top;
-static int infinity;
+goto_number *goto_map = NULL;
+static goto_number ngotos = 0;
+state_number *from_state = NULL;
+state_number *to_state = NULL;
-static void
-traverse (int i)
+/* Linked list of goto numbers. */
+typedef struct goto_list
{
- int j;
- size_t k;
- int height;
- size_t size = F (i + 1) - F(i);
+ struct goto_list *next;
+ goto_number value;
+} goto_list;
- VERTICES[++top] = i;
- INDEX[i] = height = top;
- if (R[i])
- for (j = 0; R[i][j] >= 0; ++j)
- {
- if (INDEX[R[i][j]] == 0)
- traverse (R[i][j]);
+/* LA is a LR by NTOKENS matrix of bits. LA[l, i] is 1 if the rule
+ LArule[l] is applicable in the appropriate state when the next
+ token is symbol i. If LA[l, i] and LA[l, j] are both 1 for i != j,
+ it is a conflict. */
- if (INDEX[i] > INDEX[R[i][j]])
- INDEX[i] = INDEX[R[i][j]];
+static bitsetv LA = NULL;
+size_t nLA;
- for (k = 0; k < size; ++k)
- F (i)[k] |= F (R[i][j])[k];
- }
-
- if (INDEX[i] == height)
- for (;;)
- {
- j = VERTICES[top--];
- INDEX[j] = infinity;
- if (i == j)
- break;
-
- for (k = 0; k < size; ++k)
- F (i)[k] = F (j)[k];
- }
-}
-
-
-static void
-digraph (short **relation)
-{
- int i;
-
- infinity = ngotos + 2;
- INDEX = XCALLOC (short, ngotos + 1);
- VERTICES = XCALLOC (short, ngotos + 1);
- top = 0;
-
- R = relation;
-
- for (i = 0; i < ngotos; i++)
- INDEX[i] = 0;
+/* And for the famous F variable, which name is so descriptive that a
+ comment is hardly needed. <grin>. */
+static bitsetv F = NULL;
- for (i = 0; i < ngotos; i++)
- if (INDEX[i] == 0 && R[i])
- traverse (i);
+static goto_number **includes;
+static goto_list **lookback;
- XFREE (INDEX);
- XFREE (VERTICES);
-}
-/*--------------------.
-| Build STATE_TABLE. |
-`--------------------*/
static void
-set_state_table (void)
+set_goto_map (void)
{
- /* NSTATES + 1 because lookahead for the pseudo state number NSTATES
- might be used (see conflicts.c). It is too opaque for me to
- provide a probably less hacky implementation. --akim */
- state_table = XCALLOC (state_t *, nstates + 1);
+ state_number s;
+ goto_number *temp_map;
- {
- state_t *sp;
- for (sp = first_state; sp; sp = sp->next)
- state_table[sp->number] = sp;
- }
-
- {
- shifts *sp;
- for (sp = first_shift; sp; sp = sp->next)
- assert (state_table[sp->number]->shifts == sp);
- }
+ CALLOC (goto_map, nvars + 1);
+ CALLOC (temp_map, nvars + 1);
- /* Pessimization, but simplification of the code: make sure all the
- states have a shifts, even if reduced to 0 shifts. */
- {
- int i;
- for (i = 0; i < nstates; i++)
- if (!state_table[i]->shifts)
- state_table[i]->shifts = shifts_new (0);
- }
+ ngotos = 0;
+ for (s = 0; s < nstates; ++s)
+ {
+ transitions *sp = states[s]->transitions;
+ int i;
+ for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i)
+ {
+ if (ngotos >= GOTO_NUMBER_MAXIMUM)
+ abort ();
+ ngotos++;
+ goto_map[TRANSITION_SYMBOL (sp, i) - ntokens]++;
+ }
+ }
- /* Initializing the lookaheads members. Please note that it must be
- performed after having set some of the other members which are
- used below. Change with extreme caution. */
{
+ int k = 0;
int i;
- int count = 0;
- for (i = 0; i < nstates; i++)
+ for (i = ntokens; i < nsyms; i++)
{
- int k;
- reductions *rp = state_table[i]->reductions;
- shifts *sp = state_table[i]->shifts;
-
- state_table[i]->lookaheads = count;
-
- if (rp
- && (rp->nreds > 1 || (sp->nshifts && SHIFT_IS_SHIFT (sp, 0))))
- count += rp->nreds;
- else
- state_table[i]->consistent = 1;
-
- for (k = 0; k < sp->nshifts; k++)
- if (SHIFT_IS_ERROR (sp, k))
- {
- state_table[i]->consistent = 0;
- break;
- }
+ temp_map[i - ntokens] = k;
+ k += goto_map[i - ntokens];
}
- /* Seems to be needed by conflicts.c. */
- state_table[nstates] = STATE_ALLOC (0);
- state_table[nstates]->lookaheads = count;
- }
-}
-
-
-static void
-initialize_LA (void)
-{
- int i;
- int j;
- short *np;
- reductions *rp;
-
- size_t nLA = state_table[nstates]->lookaheads;
- if (!nLA)
- nLA = 1;
-
- LA = XCALLOC (unsigned, nLA * tokensetsize);
- LAruleno = XCALLOC (short, nLA);
- lookback = XCALLOC (shorts *, nLA);
-
- np = LAruleno;
- for (i = 0; i < nstates; i++)
- if (!state_table[i]->consistent)
- if ((rp = state_table[i]->reductions))
- for (j = 0; j < rp->nreds; j++)
- *np++ = rp->rules[j];
-}
-
-
-static void
-set_goto_map (void)
-{
- shifts *sp;
- int i;
- int symbol;
- int k;
- short *temp_map;
- int state2;
- int state1;
-
- goto_map = XCALLOC (short, nvars + 1) - ntokens;
- temp_map = XCALLOC (short, nvars + 1) - ntokens;
-
- ngotos = 0;
- for (sp = first_shift; sp; sp = sp->next)
- for (i = sp->nshifts - 1; i >= 0 && SHIFT_IS_GOTO (sp, i); --i)
- {
- symbol = state_table[sp->shifts[i]]->accessing_symbol;
+ for (i = ntokens; i < nsyms; i++)
+ goto_map[i - ntokens] = temp_map[i - ntokens];
- if (ngotos == MAXSHORT)
- fatal (_("too many gotos (max %d)"), MAXSHORT);
-
- ngotos++;
- goto_map[symbol]++;
- }
-
- k = 0;
- for (i = ntokens; i < nsyms; i++)
- {
- temp_map[i] = k;
- k += goto_map[i];
- }
-
- for (i = ntokens; i < nsyms; i++)
- goto_map[i] = temp_map[i];
-
- goto_map[nsyms] = ngotos;
- temp_map[nsyms] = ngotos;
+ goto_map[nsyms - ntokens] = ngotos;
+ temp_map[nsyms - ntokens] = ngotos;
+ }
- from_state = XCALLOC (short, ngotos);
- to_state = XCALLOC (short, ngotos);
+ CALLOC (from_state, ngotos);
+ CALLOC (to_state, ngotos);
- for (sp = first_shift; sp; sp = sp->next)
+ for (s = 0; s < nstates; ++s)
{
- state1 = sp->number;
- for (i = sp->nshifts - 1; i >= 0 && SHIFT_IS_GOTO (sp, i); --i)
+ transitions *sp = states[s]->transitions;
+ int i;
+ for (i = sp->num - 1; i >= 0 && TRANSITION_IS_GOTO (sp, i); --i)
{
- state2 = sp->shifts[i];
- symbol = state_table[state2]->accessing_symbol;
-
- k = temp_map[symbol]++;
- from_state[k] = state1;
- to_state[k] = state2;
+ int k = temp_map[TRANSITION_SYMBOL (sp, i) - ntokens]++;
+ from_state[k] = s;
+ to_state[k] = sp->states[i]->number;
}
}
- XFREE (temp_map + ntokens);
+ XFREE (temp_map);
}
`----------------------------------------------------------*/
static int
-map_goto (int state, int symbol)
+map_goto (state_number s0, symbol_number sym)
{
int high;
int low;
int middle;
- int s;
+ state_number s;
- low = goto_map[symbol];
- high = goto_map[symbol + 1] - 1;
+ low = goto_map[sym - ntokens];
+ high = goto_map[sym - ntokens + 1] - 1;
- while (low <= high)
+ for (;;)
{
+ if (high < low)
+ abort ();
middle = (low + high) / 2;
s = from_state[middle];
- if (s == state)
+ if (s == s0)
return middle;
- else if (s < state)
+ else if (s < s0)
low = middle + 1;
else
high = middle - 1;
}
-
- assert (0);
- /* NOTREACHED */
- return 0;
}
static void
initialize_F (void)
{
- short **reads = XCALLOC (short *, ngotos);
- short *edge = XCALLOC (short, ngotos + 1);
+ goto_number **reads = CALLOC (reads, ngotos);
+ goto_number *edge = CALLOC (edge, ngotos + 1);
int nedges = 0;
int i;
- F = XCALLOC (unsigned, ngotos * tokensetsize);
+ F = bitsetv_create (ngotos, ntokens, BITSET_FIXED);
for (i = 0; i < ngotos; i++)
{
- int stateno = to_state[i];
- shifts *sp = state_table[stateno]->shifts;
+ state_number stateno = to_state[i];
+ transitions *sp = states[stateno]->transitions;
int j;
- for (j = 0; j < sp->nshifts && SHIFT_IS_SHIFT (sp, j); j++)
- {
- int symbol = state_table[sp->shifts[j]]->accessing_symbol;
- SETBIT (F (i), symbol);
- }
+ FOR_EACH_SHIFT (sp, j)
+ bitset_set (F[i], TRANSITION_SYMBOL (sp, j));
- for (; j < sp->nshifts; j++)
+ for (; j < sp->num; j++)
{
- int symbol = state_table[sp->shifts[j]]->accessing_symbol;
- if (nullable[symbol])
- edge[nedges++] = map_goto (stateno, symbol);
+ symbol_number sym = TRANSITION_SYMBOL (sp, j);
+ if (nullable[sym - ntokens])
+ edge[nedges++] = map_goto (stateno, sym);
}
if (nedges)
{
- reads[i] = XCALLOC (short, nedges + 1);
- shortcpy (reads[i], edge, nedges);
+ CALLOC (reads[i], nedges + 1);
+ memcpy (reads[i], edge, nedges * sizeof (edge[0]));
reads[i][nedges] = -1;
nedges = 0;
}
}
- digraph (reads);
+ relation_digraph (reads, ngotos, &F);
for (i = 0; i < ngotos; i++)
XFREE (reads[i]);
static void
-add_lookback_edge (int stateno, int ruleno, int gotono)
+add_lookback_edge (state *s, rule *r, int gotono)
{
- int i;
- int k;
- int found;
- shorts *sp;
-
- i = state_table[stateno]->lookaheads;
- k = state_table[stateno + 1]->lookaheads;
- found = 0;
- while (!found && i < k)
- {
- if (LAruleno[i] == ruleno)
- found = 1;
- else
- i++;
- }
-
- assert (found);
-
- sp = XCALLOC (shorts, 1);
- sp->next = lookback[i];
+ int ri = state_reduction_find (s, r);
+ goto_list *sp = MALLOC (sp, 1);
+ sp->next = lookback[(s->reductions->look_ahead_tokens - LA) + ri];
sp->value = gotono;
- lookback[i] = sp;
-}
-
-
-static void
-matrix_print (FILE *out, short **matrix, int n)
-{
- int i, j;
-
- for (i = 0; i < n; ++i)
- {
- fprintf (out, "%3d: ", i);
- if (matrix[i])
- for (j = 0; matrix[i][j] != -1; ++j)
- fprintf (out, "%3d ", matrix[i][j]);
- fputc ('\n', out);
- }
- fputc ('\n', out);
+ lookback[(s->reductions->look_ahead_tokens - LA) + ri] = sp;
}
-/*-------------------------------------------------------------------.
-| Return the transpose of R_ARG, of size N. Destroy R_ARG, as it is |
-| replaced with the result. |
-| |
-| R_ARG[I] is NULL or a -1 terminated list of numbers. |
-| |
-| RESULT[NUM] is NULL or the -1 terminated list of the I such as NUM |
-| is in R_ARG[I]. |
-`-------------------------------------------------------------------*/
-
-static short **
-transpose (short **R_arg, int n)
-{
- /* The result. */
- short **new_R = XCALLOC (short *, n);
- /* END_R[I] -- next entry of NEW_R[I]. */
- short **end_R = XCALLOC (short *, n);
- /* NEDGES[I] -- total size of NEW_R[I]. */
- short *nedges = XCALLOC (short, n);
- int i, j;
-
- if (trace_flag)
- {
- fputs ("transpose: input\n", stderr);
- matrix_print (stderr, R_arg, n);
- }
-
- /* Count. */
- for (i = 0; i < n; i++)
- if (R_arg[i])
- for (j = 0; R_arg[i][j] >= 0; ++j)
- ++nedges[R_arg[i][j]];
-
- /* Allocate. */
- for (i = 0; i < n; i++)
- if (nedges[i] > 0)
- {
- short *sp = XCALLOC (short, nedges[i] + 1);
- sp[nedges[i]] = -1;
- new_R[i] = sp;
- end_R[i] = sp;
- }
-
- /* Store. */
- for (i = 0; i < n; i++)
- if (R_arg[i])
- for (j = 0; R_arg[i][j] >= 0; ++j)
- {
- *end_R[R_arg[i][j]] = i;
- ++end_R[R_arg[i][j]];
- }
-
- free (nedges);
- free (end_R);
-
- /* Free the input: it is replaced with the result. */
- for (i = 0; i < n; i++)
- XFREE (R_arg[i]);
- free (R_arg);
-
- if (trace_flag)
- {
- fputs ("transpose: output\n", stderr);
- matrix_print (stderr, new_R, n);
- }
-
- return new_R;
-}
static void
build_relations (void)
{
- short *edge = XCALLOC (short, ngotos + 1);
- short *states = XCALLOC (short, ritem_longest_rhs () + 1);
+ goto_number *edge = CALLOC (edge, ngotos + 1);
+ state_number *states1 = CALLOC (states1, ritem_longest_rhs () + 1);
int i;
- includes = XCALLOC (short *, ngotos);
+ CALLOC (includes, ngotos);
for (i = 0; i < ngotos; i++)
{
int nedges = 0;
- int state1 = from_state[i];
- int symbol1 = state_table[to_state[i]]->accessing_symbol;
- short *rulep;
+ symbol_number symbol1 = states[to_state[i]]->accessing_symbol;
+ rule **rulep;
- for (rulep = derives[symbol1]; *rulep > 0; rulep++)
+ for (rulep = derives[symbol1 - ntokens]; *rulep; rulep++)
{
- int done;
+ bool done;
int length = 1;
- int stateno = state1;
- short *rp;
- states[0] = state1;
+ item_number *rp;
+ state *s = states[from_state[i]];
+ states1[0] = s->number;
- for (rp = ritem + rule_table[*rulep].rhs; *rp > 0; rp++)
+ for (rp = (*rulep)->rhs; *rp >= 0; rp++)
{
- shifts *sp = state_table[stateno]->shifts;
- int j;
- for (j = 0; j < sp->nshifts; j++)
- {
- stateno = sp->shifts[j];
- if (state_table[stateno]->accessing_symbol == *rp)
- break;
- }
-
- states[length++] = stateno;
+ s = transitions_to (s->transitions,
+ item_number_as_symbol_number (*rp));
+ states1[length++] = s->number;
}
- if (!state_table[stateno]->consistent)
- add_lookback_edge (stateno, *rulep, i);
+ if (!s->consistent)
+ add_lookback_edge (s, *rulep, i);
length--;
- done = 0;
+ done = false;
while (!done)
{
- done = 1;
+ done = true;
rp--;
/* JF added rp>=ritem && I hope to god its right! */
if (rp >= ritem && ISVAR (*rp))
{
- stateno = states[--length];
- edge[nedges++] = map_goto (stateno, *rp);
- if (nullable[*rp])
- done = 0;
+ /* Downcasting from item_number to symbol_number. */
+ edge[nedges++] = map_goto (states1[--length],
+ item_number_as_symbol_number (*rp));
+ if (nullable[*rp - ntokens])
+ done = false;
}
}
}
if (nedges)
{
int j;
- includes[i] = XCALLOC (short, nedges + 1);
+ CALLOC (includes[i], nedges + 1);
for (j = 0; j < nedges; j++)
includes[i][j] = edge[j];
includes[i][nedges] = -1;
}
XFREE (edge);
- XFREE (states);
+ XFREE (states1);
- includes = transpose (includes, ngotos);
+ relation_transpose (&includes, ngotos);
}
{
int i;
- digraph (includes);
+ relation_digraph (includes, ngotos, &F);
for (i = 0; i < ngotos; i++)
XFREE (includes[i]);
static void
-compute_lookaheads (void)
+compute_look_ahead_tokens (void)
{
- int i;
- shorts *sp;
+ size_t i;
+ goto_list *sp;
- for (i = 0; i < state_table[nstates]->lookaheads; i++)
+ for (i = 0; i < nLA; i++)
for (sp = lookback[i]; sp; sp = sp->next)
- {
- int size = LA (i + 1) - LA (i);
- int j;
- for (j = 0; j < size; ++j)
- LA (i)[j] |= F (sp->value)[j];
- }
+ bitset_or (LA[i], LA[i], F[sp->value]);
/* Free LOOKBACK. */
- for (i = 0; i < state_table[nstates]->lookaheads; i++)
- LIST_FREE (shorts, lookback[i]);
+ for (i = 0; i < nLA; i++)
+ LIST_FREE (goto_list, lookback[i]);
XFREE (lookback);
- XFREE (F);
+ bitsetv_free (F);
+}
+
+
+/*-----------------------------------------------------.
+| Count the number of look-ahead tokens required for S |
+| (N_LOOK_AHEAD_TOKENS member). |
+`-----------------------------------------------------*/
+
+static int
+state_look_ahead_tokens_count (state *s)
+{
+ int k;
+ int n_look_ahead_tokens = 0;
+ reductions *rp = s->reductions;
+ transitions *sp = s->transitions;
+
+ /* We need a look-ahead either to distinguish different
+ reductions (i.e., there are two or more), or to distinguish a
+ reduction from a shift. Otherwise, it is straightforward,
+ and the state is `consistent'. */
+ if (rp->num > 1
+ || (rp->num == 1 && sp->num &&
+ !TRANSITION_IS_DISABLED (sp, 0) && TRANSITION_IS_SHIFT (sp, 0)))
+ n_look_ahead_tokens += rp->num;
+ else
+ s->consistent = 1;
+
+ for (k = 0; k < sp->num; k++)
+ if (!TRANSITION_IS_DISABLED (sp, k) && TRANSITION_IS_ERROR (sp, k))
+ {
+ s->consistent = 0;
+ break;
+ }
+
+ return n_look_ahead_tokens;
+}
+
+
+/*-----------------------------------------------------.
+| Compute LA, NLA, and the look_ahead_tokens members. |
+`-----------------------------------------------------*/
+
+static void
+initialize_LA (void)
+{
+ state_number i;
+ bitsetv pLA;
+
+ /* Compute the total number of reductions requiring a look-ahead. */
+ nLA = 0;
+ for (i = 0; i < nstates; i++)
+ nLA += state_look_ahead_tokens_count (states[i]);
+ /* Avoid having to special case 0. */
+ if (!nLA)
+ nLA = 1;
+
+ pLA = LA = bitsetv_create (nLA, ntokens, BITSET_FIXED);
+ CALLOC (lookback, nLA);
+
+ /* Initialize the members LOOK_AHEAD_TOKENS for each state whose reductions
+ require look-ahead tokens. */
+ for (i = 0; i < nstates; i++)
+ {
+ int count = state_look_ahead_tokens_count (states[i]);
+ if (count)
+ {
+ states[i]->reductions->look_ahead_tokens = pLA;
+ pLA += count;
+ }
+ }
}
+/*----------------------------------------------.
+| Output the look-ahead tokens for each state. |
+`----------------------------------------------*/
+
+static void
+look_ahead_tokens_print (FILE *out)
+{
+ state_number i;
+ int j, k;
+ fprintf (out, "Look-ahead tokens: BEGIN\n");
+ for (i = 0; i < nstates; ++i)
+ {
+ reductions *reds = states[i]->reductions;
+ bitset_iterator iter;
+ int n_look_ahead_tokens = 0;
+
+ if (reds->look_ahead_tokens)
+ for (k = 0; k < reds->num; ++k)
+ if (reds->look_ahead_tokens[k])
+ ++n_look_ahead_tokens;
+
+ fprintf (out, "State %d: %d look-ahead tokens\n",
+ i, n_look_ahead_tokens);
+
+ if (reds->look_ahead_tokens)
+ for (j = 0; j < reds->num; ++j)
+ BITSET_FOR_EACH (iter, reds->look_ahead_tokens[j], k, 0)
+ {
+ fprintf (out, " on %d (%s) -> rule %d\n",
+ k, symbols[k]->tag,
+ reds->rules[j]->number);
+ };
+ }
+ fprintf (out, "Look-ahead tokens: END\n");
+}
+
void
lalr (void)
{
- tokensetsize = WORDSIZE (ntokens);
-
- set_state_table ();
initialize_LA ();
set_goto_map ();
initialize_F ();
build_relations ();
compute_FOLLOWS ();
- compute_lookaheads ();
+ compute_look_ahead_tokens ();
+
+ if (trace_flag & trace_sets)
+ look_ahead_tokens_print (stderr);
+}
+
+
+void
+lalr_free (void)
+{
+ state_number s;
+ for (s = 0; s < nstates; ++s)
+ states[s]->reductions->look_ahead_tokens = NULL;
+ bitsetv_free (LA);
}
projects / bison.git / blobdiff