Update.
2004-12-13  Paolo Bonzini  <bonzini@gnu.org>

	Separate parsing and creation of the NFA.  Avoided recursion on
	the (very unbalanced) parse tree.
	[BZ #611]
	* posix/regcomp.c (struct subexp_optimize, analyze_tree, calc_epsdest,
	re_dfa_add_tree_node, mark_opt_subexp_iter): Removed.
	(optimize_subexps, duplicate_tree, calc_first, calc_next,
	mark_opt_subexp): Rewritten.
	(preorder, postorder, lower_subexps, lower_subexp, link_nfa_nodes,
	create_token_tree, free_tree, free_token): New.
	(analyze): Accept a regex_t *.  Invoke the passes via the preorder and
	postorder generic visitors.  Do not initialize the fields in the
	re_dfa_t that represent the transitions.
	(free_dfa_content): Use free_token.
	(re_compile_internal): Analyze before UTF-8 optimizations.  Do not
	include optimization of subexpressions.
	(create_initial_state): Fetch the DFA node index from the first node's
	bin_tree_t *.
	(optimize_utf8): Abort on unexpected nodes, including OP_DUP_QUESTION.
	Return on COMPLEX_BRACKET.
	(duplicate_node_closure): Fix comment.
	(duplicate_node): Do not initialize the fields in the
	re_dfa_t that represent the transitions.
	(calc_eclosure, calc_inveclosure): Do not handle OP_DELETED_SUBEXP.
	(create_tree): Remove final argument.  All callers adjusted.  Rewritten
	to use create_token_tree.
	(parse_reg_exp, parse_branch, parse_expression, parse_bracket_exp,
	build_charclass_op): Use create_tree or create_token_tree instead
	of re_dfa_add_tree_node.
	(parse_dup_op): Likewise.  Also free the tree using free_tree for
	"<re>{0}", and lower OP_DUP_QUESTION to OP_ALT: "a?" is equivalent
	to "a|".  Adjust invocation of mark_opt_subexp.
	(parse_sub_exp): Create a single SUBEXP node.
	* posix/regex_internal.c (re_dfa_add_node): Remove last parameter,
	always perform as if it was 1.  Do not initialize OPT_SUBEXP and
	DUPLICATED, and initialize the DFA fields representing the transitions.
	* posix/regex_internal.h (re_dfa_add_node): Adjust prototype.
	(re_token_type_t): Move OP_DUP_PLUS and OP_DUP_QUESTION to the tokens
	section.  Add a tree-only code SUBEXP.  Remove OP_DELETED_SUBEXP.
	(bin_tree_t): Include a full re_token_t for TOKEN.  Turn FIRST and
	NEXT into pointers to trees.  Remove ECLOSURE.

2004-12-28  Paolo Bonzini  <bonzini@gnu.org >

	[BZ #605]
	* posix/regcomp.c (parse_bracket_exp): Do not modify DFA nodes
	that were already created.
	* posix/regex_internal.c (re_dfa_add_node): Set accept_mb field
	in the token if needed.
	(create_ci_newstate, create_cd_newstate): Set accept_mb field
	from the tokens' field.
	* posix/regex_internal.h (re_token_t): Add accept_mb field.
	(ACCEPT_MB_NODE): Removed.
	* posix/regexec.c (proceed_next_node, transit_states_mb,
	build_sifted_states, check_arrival_add_next_nodes): Use
	accept_mb instead of ACCEPT_MB_NODE.
This commit is contained in:
Ulrich Drepper 2005-01-26 22:42:49 +00:00
parent 629311b74a
commit 02f3550c8b
5 changed files with 583 additions and 566 deletions

View File

@ -1,3 +1,61 @@
2004-12-13 Paolo Bonzini <bonzini@gnu.org>
Separate parsing and creation of the NFA. Avoided recursion on
the (very unbalanced) parse tree.
[BZ #611]
* posix/regcomp.c (struct subexp_optimize, analyze_tree, calc_epsdest,
re_dfa_add_tree_node, mark_opt_subexp_iter): Removed.
(optimize_subexps, duplicate_tree, calc_first, calc_next,
mark_opt_subexp): Rewritten.
(preorder, postorder, lower_subexps, lower_subexp, link_nfa_nodes,
create_token_tree, free_tree, free_token): New.
(analyze): Accept a regex_t *. Invoke the passes via the preorder and
postorder generic visitors. Do not initialize the fields in the
re_dfa_t that represent the transitions.
(free_dfa_content): Use free_token.
(re_compile_internal): Analyze before UTF-8 optimizations. Do not
include optimization of subexpressions.
(create_initial_state): Fetch the DFA node index from the first node's
bin_tree_t *.
(optimize_utf8): Abort on unexpected nodes, including OP_DUP_QUESTION.
Return on COMPLEX_BRACKET.
(duplicate_node_closure): Fix comment.
(duplicate_node): Do not initialize the fields in the
re_dfa_t that represent the transitions.
(calc_eclosure, calc_inveclosure): Do not handle OP_DELETED_SUBEXP.
(create_tree): Remove final argument. All callers adjusted. Rewritten
to use create_token_tree.
(parse_reg_exp, parse_branch, parse_expression, parse_bracket_exp,
build_charclass_op): Use create_tree or create_token_tree instead
of re_dfa_add_tree_node.
(parse_dup_op): Likewise. Also free the tree using free_tree for
"<re>{0}", and lower OP_DUP_QUESTION to OP_ALT: "a?" is equivalent
to "a|". Adjust invocation of mark_opt_subexp.
(parse_sub_exp): Create a single SUBEXP node.
* posix/regex_internal.c (re_dfa_add_node): Remove last parameter,
always perform as if it was 1. Do not initialize OPT_SUBEXP and
DUPLICATED, and initialize the DFA fields representing the transitions.
* posix/regex_internal.h (re_dfa_add_node): Adjust prototype.
(re_token_type_t): Move OP_DUP_PLUS and OP_DUP_QUESTION to the tokens
section. Add a tree-only code SUBEXP. Remove OP_DELETED_SUBEXP.
(bin_tree_t): Include a full re_token_t for TOKEN. Turn FIRST and
NEXT into pointers to trees. Remove ECLOSURE.
2004-12-28 Paolo Bonzini <bonzini@gnu.org >
[BZ #605]
* posix/regcomp.c (parse_bracket_exp): Do not modify DFA nodes
that were already created.
* posix/regex_internal.c (re_dfa_add_node): Set accept_mb field
in the token if needed.
(create_ci_newstate, create_cd_newstate): Set accept_mb field
from the tokens' field.
* posix/regex_internal.h (re_token_t): Add accept_mb field.
(ACCEPT_MB_NODE): Removed.
* posix/regexec.c (proceed_next_node, transit_states_mb,
build_sifted_states, check_arrival_add_next_nodes): Use
accept_mb instead of ACCEPT_MB_NODE.
2005-01-26 Ulrich Drepper <drepper@redhat.com>
* debug/chk_fail.c (__chk_fail): Print program name in final message.

File diff suppressed because it is too large Load Diff

View File

@ -1330,29 +1330,26 @@ re_node_set_remove_at (set, idx)
Or return -1, if an error will be occured. */
static int
re_dfa_add_node (dfa, token, mode)
re_dfa_add_node (dfa, token)
re_dfa_t *dfa;
re_token_t token;
int mode;
{
int type = token.type;
if (BE (dfa->nodes_len >= dfa->nodes_alloc, 0))
{
int new_nodes_alloc = dfa->nodes_alloc * 2;
int *new_nexts, *new_indices;
re_node_set *new_edests, *new_eclosures, *new_inveclosures;
re_token_t *new_array = re_realloc (dfa->nodes, re_token_t,
new_nodes_alloc);
if (BE (new_array == NULL, 0))
return -1;
dfa->nodes = new_array;
if (mode)
{
int *new_nexts, *new_indices;
re_node_set *new_edests, *new_eclosures, *new_inveclosures;
new_nexts = re_realloc (dfa->nexts, int, new_nodes_alloc);
new_indices = re_realloc (dfa->org_indices, int, new_nodes_alloc);
new_edests = re_realloc (dfa->edests, re_node_set, new_nodes_alloc);
new_eclosures = re_realloc (dfa->eclosures, re_node_set,
new_nodes_alloc);
new_eclosures = re_realloc (dfa->eclosures, re_node_set, new_nodes_alloc);
new_inveclosures = re_realloc (dfa->inveclosures, re_node_set,
new_nodes_alloc);
if (BE (new_nexts == NULL || new_indices == NULL
@ -1364,13 +1361,18 @@ re_dfa_add_node (dfa, token, mode)
dfa->edests = new_edests;
dfa->eclosures = new_eclosures;
dfa->inveclosures = new_inveclosures;
}
dfa->nodes_alloc = new_nodes_alloc;
}
dfa->nodes[dfa->nodes_len] = token;
dfa->nodes[dfa->nodes_len].opt_subexp = 0;
dfa->nodes[dfa->nodes_len].duplicated = 0;
dfa->nodes[dfa->nodes_len].constraint = 0;
#ifdef RE_ENABLE_I18N
dfa->nodes[dfa->nodes_len].accept_mb =
(type == OP_PERIOD && dfa->mb_cur_max > 1) || type == COMPLEX_BRACKET;
#endif
dfa->nexts[dfa->nodes_len] = -1;
re_node_set_init_empty (dfa->edests + dfa->nodes_len);
re_node_set_init_empty (dfa->eclosures + dfa->nodes_len);
re_node_set_init_empty (dfa->inveclosures + dfa->nodes_len);
return dfa->nodes_len++;
}
@ -1551,16 +1553,13 @@ create_ci_newstate (dfa, nodes, hash)
re_token_type_t type = node->type;
if (type == CHARACTER && !node->constraint)
continue;
#ifdef RE_ENABLE_I18N
newstate->accept_mb |= node->accept_mb;
#endif /* RE_ENABLE_I18N */
/* If the state has the halt node, the state is a halt state. */
else if (type == END_OF_RE)
if (type == END_OF_RE)
newstate->halt = 1;
#ifdef RE_ENABLE_I18N
else if (type == COMPLEX_BRACKET
|| type == OP_UTF8_PERIOD
|| (type == OP_PERIOD && dfa->mb_cur_max > 1))
newstate->accept_mb = 1;
#endif /* RE_ENABLE_I18N */
else if (type == OP_BACK_REF)
newstate->has_backref = 1;
else if (type == ANCHOR || node->constraint)
@ -1611,15 +1610,13 @@ create_cd_newstate (dfa, nodes, context, hash)
if (type == CHARACTER && !constraint)
continue;
/* If the state has the halt node, the state is a halt state. */
else if (type == END_OF_RE)
newstate->halt = 1;
#ifdef RE_ENABLE_I18N
else if (type == COMPLEX_BRACKET
|| type == OP_UTF8_PERIOD
|| (type == OP_PERIOD && dfa->mb_cur_max > 1))
newstate->accept_mb = 1;
newstate->accept_mb |= node->accept_mb;
#endif /* RE_ENABLE_I18N */
/* If the state has the halt node, the state is a halt state. */
if (type == END_OF_RE)
newstate->halt = 1;
else if (type == OP_BACK_REF)
newstate->has_backref = 1;
else if (type == ANCHOR)

View File

@ -189,16 +189,16 @@ typedef enum
OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
OP_ALT = EPSILON_BIT | 2,
OP_DUP_ASTERISK = EPSILON_BIT | 3,
OP_DUP_PLUS = EPSILON_BIT | 4,
OP_DUP_QUESTION = EPSILON_BIT | 5,
ANCHOR = EPSILON_BIT | 6,
OP_DELETED_SUBEXP = EPSILON_BIT | 7,
ANCHOR = EPSILON_BIT | 4,
/* Tree type, these are used only by tree. */
CONCAT = 16,
SUBEXP = 17,
/* Token type, these are used only by token. */
OP_OPEN_BRACKET = 17,
OP_DUP_PLUS = 18,
OP_DUP_QUESTION,
OP_OPEN_BRACKET,
OP_CLOSE_BRACKET,
OP_CHARSET_RANGE,
OP_OPEN_DUP_NUM,
@ -287,6 +287,7 @@ typedef struct
unsigned int duplicated : 1;
unsigned int opt_subexp : 1;
#ifdef RE_ENABLE_I18N
unsigned int accept_mb : 1;
/* These 2 bits can be moved into the union if needed (e.g. if running out
of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
unsigned int mb_partial : 1;
@ -295,8 +296,6 @@ typedef struct
} re_token_t;
#define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
#define ACCEPT_MB_NODE(type) \
((type) >= OP_PERIOD && (type) <= OP_UTF8_PERIOD)
struct re_string_t
{
@ -432,15 +431,14 @@ struct bin_tree_t
struct bin_tree_t *parent;
struct bin_tree_t *left;
struct bin_tree_t *right;
struct bin_tree_t *first;
struct bin_tree_t *next;
re_token_t token;
/* `node_idx' is the index in dfa->nodes, if `type' == 0.
Otherwise `type' indicate the type of this node. */
re_token_type_t type;
int node_idx;
int first;
int next;
re_node_set eclosure;
};
typedef struct bin_tree_t bin_tree_t;
@ -680,7 +678,7 @@ static void re_node_set_remove_at (re_node_set *set, int idx) internal_function;
(re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
#define re_node_set_empty(p) ((p)->nelem = 0)
#define re_node_set_free(set) re_free ((set)->elems)
static int re_dfa_add_node (re_dfa_t *dfa, re_token_t token, int mode) internal_function;
static int re_dfa_add_node (re_dfa_t *dfa, re_token_t token) internal_function;
static re_dfastate_t *re_acquire_state (reg_errcode_t *err, re_dfa_t *dfa,
const re_node_set *nodes) internal_function;
static re_dfastate_t *re_acquire_state_context (reg_errcode_t *err,

View File

@ -1262,7 +1262,7 @@ proceed_next_node (mctx, nregs, regs, pidx, node, eps_via_nodes, fs)
re_token_type_t type = dfa->nodes[node].type;
#ifdef RE_ENABLE_I18N
if (ACCEPT_MB_NODE (type))
if (dfa->nodes[node].accept_mb)
naccepted = check_node_accept_bytes (dfa, node, &mctx->input, *pidx);
else
#endif /* RE_ENABLE_I18N */
@ -1624,15 +1624,13 @@ build_sifted_states (mctx, sctx, str_idx, cur_dest)
int naccepted = 0;
int ret;
#if defined DEBUG || defined RE_ENABLE_I18N
re_token_type_t type = dfa->nodes[prev_node].type;
#endif
#ifdef DEBUG
re_token_type_t type = dfa->nodes[prev_node].type;
assert (!IS_EPSILON_NODE (type));
#endif
#ifdef RE_ENABLE_I18N
/* If the node may accept `multi byte'. */
if (ACCEPT_MB_NODE (type))
if (dfa->nodes[prev_node].accept_mb)
naccepted = sift_states_iter_mb (mctx, sctx, prev_node,
str_idx, sctx->last_str_idx);
#endif /* RE_ENABLE_I18N */
@ -2486,7 +2484,7 @@ transit_state_mb (mctx, pstate)
}
/* How many bytes the node can accept? */
if (ACCEPT_MB_NODE (dfa->nodes[cur_node_idx].type))
if (dfa->nodes[cur_node_idx].accept_mb)
naccepted = check_node_accept_bytes (dfa, cur_node_idx, &mctx->input,
re_string_cur_idx (&mctx->input));
if (naccepted == 0)
@ -3020,15 +3018,13 @@ check_arrival_add_next_nodes (mctx, str_idx, cur_nodes, next_nodes)
{
int naccepted = 0;
int cur_node = cur_nodes->elems[cur_idx];
#if defined DEBUG || defined RE_ENABLE_I18N
re_token_type_t type = dfa->nodes[cur_node].type;
#endif
#ifdef DEBUG
re_token_type_t type = dfa->nodes[cur_node].type;
assert (!IS_EPSILON_NODE (type));
#endif
#ifdef RE_ENABLE_I18N
/* If the node may accept `multi byte'. */
if (ACCEPT_MB_NODE (type))
if (dfa->nodes[cur_node].accept_mb)
{
naccepted = check_node_accept_bytes (dfa, cur_node, &mctx->input,
str_idx);