[PATCH] Don't pass preg around unnecessarily
Jakub Jelinek
jakub@redhat.com
Fri Jan 2 17:21:00 GMT 2004
Hi!
This patch on top of
http://sources.redhat.com/ml/libc-hacker/2004-01/msg00000.html
changes most of the const regex_t *preg passing around into passing
re_dfa_t *dfa directly (with the exception of re_search_internal
and its callers and set_regs). These functions were mainly interested
just in preg->buffer (ie. dfa which is now passed around directly)
and only rarely in preg->syntax (which makes sense to put into
re_dfa_t).
2004-01-02 Jakub Jelinek <jakub@redhat.com>
* posix/regex_internal.h (re_dfa_t): Add syntax field.
* posix/regcomp.c (parse): Initialize dfa->syntax.
* posix/regexec.c (acquire_init_state_context,
prune_impossible_nodes, check_matching, check_halt_state_context,
proceed_next_node, sift_states_iter_mb, sift_states_backward,
update_cur_sifted_state, sift_states_bkref, transit_state,
transit_state_sb, transit_state_mb, transit_state_bkref,
get_subexp, get_subexp_sub, check_arrival, expand_bkref_cache,
build_trtable): Remove preg argument, add dfa argument instead
and remove dfa = preg->buffer initialization in the body.
Adjust all callers.
(check_node_accept_bytes, group_nodes_into_DFAstates,
check_node_accept): Likewise. Use dfa->syntax instead of
preg->syntax.
(check_arrival_add_next_nodes): Remove preg argument.
--- libc/posix/regex_internal.h.jj 2004-01-02 16:17:11.000000000 +0100
+++ libc/posix/regex_internal.h 2004-01-02 17:20:23.000000000 +0100
@@ -636,6 +636,7 @@ struct re_dfa_t
unsigned int word_ops_used : 1;
int mb_cur_max;
bitset word_char;
+ reg_syntax_t syntax;
#ifdef DEBUG
char* re_str;
#endif
--- libc/posix/regcomp.c.jj 2004-01-02 15:42:56.000000000 +0100
+++ libc/posix/regcomp.c 2004-01-02 17:23:18.000000000 +0100
@@ -1955,6 +1955,7 @@ parse (regexp, preg, syntax, err)
re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
bin_tree_t *tree, *eor, *root;
re_token_t current_token;
+ dfa->syntax = syntax;
fetch_token (¤t_token, regexp, syntax | RE_CARET_ANCHORS_HERE);
tree = parse_reg_exp (regexp, preg, ¤t_token, syntax, 0, err);
if (BE (*err != REG_NOERROR && tree == NULL, 0))
--- libc/posix/regexec.c.jj 2004-01-02 16:17:11.000000000 +0100
+++ libc/posix/regexec.c 2004-01-02 17:29:41.000000000 +0100
@@ -51,21 +51,21 @@ static int re_search_stub (struct re_pat
static unsigned re_copy_regs (struct re_registers *regs, regmatch_t *pmatch,
int nregs, int regs_allocated) internal_function;
static inline re_dfastate_t *acquire_init_state_context
- (reg_errcode_t *err, const regex_t *preg, const re_match_context_t *mctx,
+ (reg_errcode_t *err, re_dfa_t *dfa, const re_match_context_t *mctx,
int idx) __attribute ((always_inline)) internal_function;
-static reg_errcode_t prune_impossible_nodes (const regex_t *preg,
+static reg_errcode_t prune_impossible_nodes (re_dfa_t *dfa,
re_match_context_t *mctx) internal_function;
-static int check_matching (const regex_t *preg, re_match_context_t *mctx,
+static int check_matching (re_dfa_t *dfa, re_match_context_t *mctx,
int fl_longest_match) internal_function;
static int check_halt_node_context (const re_dfa_t *dfa, int node,
unsigned int context) internal_function;
-static int check_halt_state_context (const regex_t *preg,
+static int check_halt_state_context (re_dfa_t *dfa,
const re_dfastate_t *state,
const re_match_context_t *mctx, int idx) internal_function;
static void update_regs (re_dfa_t *dfa, regmatch_t *pmatch,
regmatch_t *prev_idx_match, int cur_node,
int cur_idx, int nmatch) internal_function;
-static int proceed_next_node (const regex_t *preg, int nregs, regmatch_t *regs,
+static int proceed_next_node (re_dfa_t *dfa, int nregs, regmatch_t *regs,
const re_match_context_t *mctx,
int *pidx, int node, re_node_set *eps_via_nodes,
struct re_fail_stack_t *fs) internal_function;
@@ -82,15 +82,15 @@ static reg_errcode_t set_regs (const reg
static reg_errcode_t free_fail_stack_return (struct re_fail_stack_t *fs) internal_function;
#ifdef RE_ENABLE_I18N
-static int sift_states_iter_mb (const regex_t *preg,
+static int sift_states_iter_mb (re_dfa_t *dfa,
const re_match_context_t *mctx,
re_sift_context_t *sctx,
int node_idx, int str_idx, int max_str_idx) internal_function;
#endif /* RE_ENABLE_I18N */
-static reg_errcode_t sift_states_backward (const regex_t *preg,
+static reg_errcode_t sift_states_backward (re_dfa_t *dfa,
re_match_context_t *mctx,
re_sift_context_t *sctx) internal_function;
-static reg_errcode_t update_cur_sifted_state (const regex_t *preg,
+static reg_errcode_t update_cur_sifted_state (re_dfa_t *dfa,
re_match_context_t *mctx,
re_sift_context_t *sctx,
int str_idx,
@@ -113,7 +113,7 @@ static reg_errcode_t check_subexp_limits
re_node_set *limits,
struct re_backref_cache_entry *bkref_ents,
int str_idx) internal_function;
-static reg_errcode_t sift_states_bkref (const regex_t *preg,
+static reg_errcode_t sift_states_bkref (re_dfa_t *dfa,
re_match_context_t *mctx,
re_sift_context_t *sctx,
int str_idx, re_node_set *dest_nodes) internal_function;
@@ -121,7 +121,7 @@ static reg_errcode_t clean_state_log_if_
int next_state_log_idx) internal_function;
static reg_errcode_t merge_state_array (re_dfa_t *dfa, re_dfastate_t **dst,
re_dfastate_t **src, int num) internal_function;
-static re_dfastate_t *transit_state (reg_errcode_t *err, const regex_t *preg,
+static re_dfastate_t *transit_state (reg_errcode_t *err, re_dfa_t *dfa,
re_match_context_t *mctx,
re_dfastate_t *state) internal_function;
static reg_errcode_t check_subexp_matching_top (re_dfa_t *dfa,
@@ -129,34 +129,33 @@ static reg_errcode_t check_subexp_matchi
re_node_set *cur_nodes,
int str_idx) internal_function;
#if 0
-static re_dfastate_t *transit_state_sb (reg_errcode_t *err, const regex_t *preg,
+static re_dfastate_t *transit_state_sb (reg_errcode_t *err, re_dfa_t *dfa,
re_dfastate_t *pstate,
re_match_context_t *mctx) internal_function;
#endif
#ifdef RE_ENABLE_I18N
-static reg_errcode_t transit_state_mb (const regex_t *preg,
+static reg_errcode_t transit_state_mb (re_dfa_t *dfa,
re_dfastate_t *pstate,
re_match_context_t *mctx) internal_function;
#endif /* RE_ENABLE_I18N */
-static reg_errcode_t transit_state_bkref (const regex_t *preg,
+static reg_errcode_t transit_state_bkref (re_dfa_t *dfa,
const re_node_set *nodes,
re_match_context_t *mctx) internal_function;
-static reg_errcode_t get_subexp (const regex_t *preg, re_match_context_t *mctx,
+static reg_errcode_t get_subexp (re_dfa_t *dfa, re_match_context_t *mctx,
int bkref_node, int bkref_str_idx) internal_function;
-static reg_errcode_t get_subexp_sub (const regex_t *preg,
+static reg_errcode_t get_subexp_sub (re_dfa_t *dfa,
re_match_context_t *mctx,
const re_sub_match_top_t *sub_top,
re_sub_match_last_t *sub_last,
int bkref_node, int bkref_str) internal_function;
static int find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes,
int subexp_idx, int type) internal_function;
-static reg_errcode_t check_arrival (const regex_t *preg,
+static reg_errcode_t check_arrival (re_dfa_t *dfa,
re_match_context_t *mctx,
state_array_t *path, int top_node,
int top_str, int last_node, int last_str,
int type) internal_function;
-static reg_errcode_t check_arrival_add_next_nodes (const regex_t *preg,
- re_dfa_t *dfa,
+static reg_errcode_t check_arrival_add_next_nodes (re_dfa_t *dfa,
re_match_context_t *mctx,
int str_idx,
re_node_set *cur_nodes,
@@ -168,26 +167,26 @@ static reg_errcode_t check_arrival_expan
re_node_set *dst_nodes,
int target, int ex_subexp,
int type) internal_function;
-static reg_errcode_t expand_bkref_cache (const regex_t *preg,
+static reg_errcode_t expand_bkref_cache (re_dfa_t *dfa,
re_match_context_t *mctx,
re_node_set *cur_nodes, int cur_str,
int last_str, int subexp_num,
int type) internal_function;
-static re_dfastate_t **build_trtable (const regex_t *dfa,
+static re_dfastate_t **build_trtable (re_dfa_t *dfa,
re_dfastate_t *state) internal_function;
#ifdef RE_ENABLE_I18N
-static int check_node_accept_bytes (const regex_t *preg, int node_idx,
+static int check_node_accept_bytes (re_dfa_t *dfa, int node_idx,
const re_string_t *input, int idx) internal_function;
# ifdef _LIBC
static unsigned int find_collation_sequence_value (const unsigned char *mbs,
size_t name_len) internal_function;
# endif /* _LIBC */
#endif /* RE_ENABLE_I18N */
-static int group_nodes_into_DFAstates (const regex_t *dfa,
+static int group_nodes_into_DFAstates (re_dfa_t *dfa,
const re_dfastate_t *state,
re_node_set *states_node,
bitset *states_ch) internal_function;
-static int check_node_accept (const regex_t *preg, const re_token_t *node,
+static int check_node_accept (re_dfa_t *dfa, const re_token_t *node,
const re_match_context_t *mctx, int idx) internal_function;
static reg_errcode_t extend_buffers (re_match_context_t *mctx) internal_function;
@@ -741,7 +740,7 @@ re_search_internal (preg, string, length
/* It seems to be appropriate one, then use the matcher. */
/* We assume that the matching starts from 0. */
mctx.state_log_top = mctx.nbkref_ents = mctx.max_mb_elem_len = 0;
- match_last = check_matching (preg, &mctx, fl_longest_match);
+ match_last = check_matching (dfa, &mctx, fl_longest_match);
if (match_last != -1)
{
if (BE (match_last == -2, 0))
@@ -755,13 +754,13 @@ re_search_internal (preg, string, length
if ((!preg->no_sub && nmatch > 1) || dfa->nbackref)
{
re_dfastate_t *pstate = mctx.state_log[match_last];
- mctx.last_node = check_halt_state_context (preg, pstate,
+ mctx.last_node = check_halt_state_context (dfa, pstate,
&mctx, match_last);
}
if ((!preg->no_sub && nmatch > 1 && dfa->has_plural_match)
|| dfa->nbackref)
{
- err = prune_impossible_nodes (preg, &mctx);
+ err = prune_impossible_nodes (dfa, &mctx);
if (err == REG_NOERROR)
break;
if (BE (err != REG_NOMATCH, 0))
@@ -836,13 +835,12 @@ re_search_internal (preg, string, length
}
static reg_errcode_t
-prune_impossible_nodes (preg, mctx)
- const regex_t *preg;
+prune_impossible_nodes (dfa, mctx)
+ re_dfa_t *dfa;
re_match_context_t *mctx;
{
int halt_node, match_last;
reg_errcode_t ret;
- re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
re_dfastate_t **sifted_states;
re_dfastate_t **lim_states = NULL;
re_sift_context_t sctx;
@@ -872,7 +870,7 @@ prune_impossible_nodes (preg, mctx)
match_ctx_clear_flag (mctx);
sift_ctx_init (&sctx, sifted_states, lim_states, halt_node,
match_last, 0);
- ret = sift_states_backward (preg, mctx, &sctx);
+ ret = sift_states_backward (dfa, mctx, &sctx);
re_node_set_free (&sctx.limits);
if (BE (ret != REG_NOERROR, 0))
goto free_return;
@@ -888,7 +886,7 @@ prune_impossible_nodes (preg, mctx)
}
} while (mctx->state_log[match_last] == NULL
|| !mctx->state_log[match_last]->halt);
- halt_node = check_halt_state_context (preg,
+ halt_node = check_halt_state_context (dfa,
mctx->state_log[match_last],
mctx, match_last);
}
@@ -903,7 +901,7 @@ prune_impossible_nodes (preg, mctx)
{
sift_ctx_init (&sctx, sifted_states, lim_states, halt_node,
match_last, 0);
- ret = sift_states_backward (preg, mctx, &sctx);
+ ret = sift_states_backward (dfa, mctx, &sctx);
re_node_set_free (&sctx.limits);
if (BE (ret != REG_NOERROR, 0))
goto free_return;
@@ -925,14 +923,12 @@ prune_impossible_nodes (preg, mctx)
since initial states may have constraints like "\<", "^", etc.. */
static inline re_dfastate_t *
-acquire_init_state_context (err, preg, mctx, idx)
+acquire_init_state_context (err, dfa, mctx, idx)
reg_errcode_t *err;
- const regex_t *preg;
+ re_dfa_t *dfa;
const re_match_context_t *mctx;
int idx;
{
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
-
*err = REG_NOERROR;
if (dfa->init_state->has_constraint)
{
@@ -969,19 +965,18 @@ acquire_init_state_context (err, preg, m
index of the buffer. */
static int
-check_matching (preg, mctx, fl_longest_match)
- const regex_t *preg;
+check_matching (dfa, mctx, fl_longest_match)
+ re_dfa_t *dfa;
re_match_context_t *mctx;
int fl_longest_match;
{
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
reg_errcode_t err;
int match = 0;
int match_last = -1;
int cur_str_idx = re_string_cur_idx (&mctx->input);
re_dfastate_t *cur_state;
- cur_state = acquire_init_state_context (&err, preg, mctx, cur_str_idx);
+ cur_state = acquire_init_state_context (&err, dfa, mctx, cur_str_idx);
/* An initial state must not be NULL(invalid state). */
if (BE (cur_state == NULL, 0))
return -2;
@@ -998,7 +993,7 @@ check_matching (preg, mctx, fl_longest_m
if (cur_state->has_backref)
{
- err = transit_state_bkref (preg, &cur_state->nodes, mctx);
+ err = transit_state_bkref (dfa, &cur_state->nodes, mctx);
if (BE (err != REG_NOERROR, 0))
return err;
}
@@ -1008,7 +1003,7 @@ check_matching (preg, mctx, fl_longest_m
if (BE (cur_state->halt, 0))
{
if (!cur_state->has_constraint
- || check_halt_state_context (preg, cur_state, mctx, cur_str_idx))
+ || check_halt_state_context (dfa, cur_state, mctx, cur_str_idx))
{
if (!fl_longest_match)
return cur_str_idx;
@@ -1022,7 +1017,7 @@ check_matching (preg, mctx, fl_longest_m
while (!re_string_eoi (&mctx->input))
{
- cur_state = transit_state (&err, preg, mctx, cur_state);
+ cur_state = transit_state (&err, dfa, mctx, cur_state);
if (cur_state == NULL) /* Reached at the invalid state or an error. */
{
cur_str_idx = re_string_cur_idx (&mctx->input);
@@ -1051,7 +1046,7 @@ check_matching (preg, mctx, fl_longest_m
/* Reached at a halt state.
Check the halt state can satisfy the current context. */
if (!cur_state->has_constraint
- || check_halt_state_context (preg, cur_state, mctx,
+ || check_halt_state_context (dfa, cur_state, mctx,
re_string_cur_idx (&mctx->input)))
{
/* We found an appropriate halt state. */
@@ -1088,13 +1083,12 @@ static int check_halt_node_context (dfa,
match the context, return the node. */
static int
-check_halt_state_context (preg, state, mctx, idx)
- const regex_t *preg;
+check_halt_state_context (dfa, state, mctx, idx)
+ re_dfa_t *dfa;
const re_dfastate_t *state;
const re_match_context_t *mctx;
int idx;
{
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
int i;
unsigned int context;
#ifdef DEBUG
@@ -1113,15 +1107,14 @@ check_halt_state_context (preg, state, m
of errors. */
static int
-proceed_next_node (preg, nregs, regs, mctx, pidx, node, eps_via_nodes, fs)
- const regex_t *preg;
+proceed_next_node (dfa, nregs, regs, mctx, pidx, node, eps_via_nodes, fs)
+ re_dfa_t *dfa;
regmatch_t *regs;
const re_match_context_t *mctx;
int nregs, *pidx, node;
re_node_set *eps_via_nodes;
struct re_fail_stack_t *fs;
{
- re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
int i, err, dest_node;
dest_node = -1;
if (IS_EPSILON_NODE (dfa->nodes[node].type))
@@ -1159,7 +1152,7 @@ proceed_next_node (preg, nregs, regs, mc
#ifdef RE_ENABLE_I18N
if (ACCEPT_MB_NODE (type))
- naccepted = check_node_accept_bytes (preg, node, &mctx->input, *pidx);
+ naccepted = check_node_accept_bytes (dfa, node, &mctx->input, *pidx);
else
#endif /* RE_ENABLE_I18N */
if (type == OP_BACK_REF)
@@ -1192,7 +1185,7 @@ proceed_next_node (preg, nregs, regs, mc
}
if (naccepted != 0
- || check_node_accept (preg, dfa->nodes + node, mctx, *pidx))
+ || check_node_accept (dfa, dfa->nodes + node, mctx, *pidx))
{
dest_node = dfa->nexts[node];
*pidx = (naccepted == 0) ? *pidx + 1 : *pidx + naccepted;
@@ -1322,7 +1315,7 @@ set_regs (preg, mctx, nmatch, pmatch, fl
}
/* Proceed to next node. */
- cur_node = proceed_next_node (preg, nmatch, pmatch, mctx, &idx, cur_node,
+ cur_node = proceed_next_node (dfa, nmatch, pmatch, mctx, &idx, cur_node,
&eps_via_nodes, fs);
if (BE (cur_node < 0, 0))
@@ -1438,13 +1431,12 @@ update_regs (dfa, pmatch, prev_idx_match
((state) != NULL && re_node_set_contains (&(state)->nodes, node))
static reg_errcode_t
-sift_states_backward (preg, mctx, sctx)
- const regex_t *preg;
+sift_states_backward (dfa, mctx, sctx)
+ re_dfa_t *dfa;
re_match_context_t *mctx;
re_sift_context_t *sctx;
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
int null_cnt = 0;
int str_idx = sctx->last_str_idx;
re_node_set cur_dest;
@@ -1460,7 +1452,7 @@ sift_states_backward (preg, mctx, sctx)
err = re_node_set_init_1 (&cur_dest, sctx->last_node);
if (BE (err != REG_NOERROR, 0))
return err;
- err = update_cur_sifted_state (preg, mctx, sctx, str_idx, &cur_dest);
+ err = update_cur_sifted_state (dfa, mctx, sctx, str_idx, &cur_dest);
if (BE (err != REG_NOERROR, 0))
goto free_return;
@@ -1499,7 +1491,7 @@ sift_states_backward (preg, mctx, sctx)
#ifdef RE_ENABLE_I18N
/* If the node may accept `multi byte'. */
if (ACCEPT_MB_NODE (type))
- naccepted = sift_states_iter_mb (preg, mctx, sctx, prev_node,
+ naccepted = sift_states_iter_mb (dfa, mctx, sctx, prev_node,
str_idx, sctx->last_str_idx);
#endif /* RE_ENABLE_I18N */
@@ -1507,7 +1499,7 @@ sift_states_backward (preg, mctx, sctx)
See update_cur_sifted_state(). */
if (!naccepted
- && check_node_accept (preg, dfa->nodes + prev_node, mctx,
+ && check_node_accept (dfa, dfa->nodes + prev_node, mctx,
str_idx)
&& STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + 1],
dfa->nexts[prev_node]))
@@ -1536,7 +1528,7 @@ sift_states_backward (preg, mctx, sctx)
- It can epsilon transit to a node in CUR_DEST.
- It is in CUR_SRC.
And update state_log. */
- err = update_cur_sifted_state (preg, mctx, sctx, str_idx, &cur_dest);
+ err = update_cur_sifted_state (dfa, mctx, sctx, str_idx, &cur_dest);
if (BE (err != REG_NOERROR, 0))
goto free_return;
}
@@ -1604,15 +1596,14 @@ merge_state_array (dfa, dst, src, num)
}
static reg_errcode_t
-update_cur_sifted_state (preg, mctx, sctx, str_idx, dest_nodes)
- const regex_t *preg;
+update_cur_sifted_state (dfa, mctx, sctx, str_idx, dest_nodes)
+ re_dfa_t *dfa;
re_match_context_t *mctx;
re_sift_context_t *sctx;
int str_idx;
re_node_set *dest_nodes;
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
const re_node_set *candidates;
candidates = ((mctx->state_log[str_idx] == NULL) ? &empty_set
: &mctx->state_log[str_idx]->nodes);
@@ -1643,7 +1634,7 @@ update_cur_sifted_state (preg, mctx, sct
if ((mctx->state_log[str_idx] != NULL
&& mctx->state_log[str_idx]->has_backref))
{
- err = sift_states_bkref (preg, mctx, sctx, str_idx, dest_nodes);
+ err = sift_states_bkref (dfa, mctx, sctx, str_idx, dest_nodes);
if (BE (err != REG_NOERROR, 0))
return err;
}
@@ -1958,15 +1949,14 @@ check_subexp_limits (dfa, dest_nodes, ca
}
static reg_errcode_t
-sift_states_bkref (preg, mctx, sctx, str_idx, dest_nodes)
- const regex_t *preg;
+sift_states_bkref (dfa, mctx, sctx, str_idx, dest_nodes)
+ re_dfa_t *dfa;
re_match_context_t *mctx;
re_sift_context_t *sctx;
int str_idx;
re_node_set *dest_nodes;
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
int node_idx, node;
re_sift_context_t local_sctx;
const re_node_set *candidates;
@@ -2039,7 +2029,7 @@ sift_states_bkref (preg, mctx, sctx, str
goto free_return;
}
cur_state = local_sctx.sifted_states[str_idx];
- err = sift_states_backward (preg, mctx, &local_sctx);
+ err = sift_states_backward (dfa, mctx, &local_sctx);
if (BE (err != REG_NOERROR, 0))
goto free_return;
if (sctx->limited_states != NULL)
@@ -2082,16 +2072,15 @@ sift_states_bkref (preg, mctx, sctx, str
#ifdef RE_ENABLE_I18N
static int
-sift_states_iter_mb (preg, mctx, sctx, node_idx, str_idx, max_str_idx)
- const regex_t *preg;
+sift_states_iter_mb (dfa, mctx, sctx, node_idx, str_idx, max_str_idx)
+ re_dfa_t *dfa;
const re_match_context_t *mctx;
re_sift_context_t *sctx;
int node_idx, str_idx, max_str_idx;
{
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
int naccepted;
/* Check the node can accept `multi byte'. */
- naccepted = check_node_accept_bytes (preg, node_idx, &mctx->input, str_idx);
+ naccepted = check_node_accept_bytes (dfa, node_idx, &mctx->input, str_idx);
if (naccepted > 0 && str_idx + naccepted <= max_str_idx &&
!STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + naccepted],
dfa->nexts[node_idx]))
@@ -2114,13 +2103,12 @@ sift_states_iter_mb (preg, mctx, sctx, n
update the destination of STATE_LOG. */
static re_dfastate_t *
-transit_state (err, preg, mctx, state)
+transit_state (err, dfa, mctx, state)
reg_errcode_t *err;
- const regex_t *preg;
+ re_dfa_t *dfa;
re_match_context_t *mctx;
re_dfastate_t *state;
{
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
re_dfastate_t **trtable, *next_state;
unsigned char ch;
int cur_idx;
@@ -2146,7 +2134,7 @@ transit_state (err, preg, mctx, state)
/* If the current state can accept multibyte. */
if (state->accept_mb)
{
- *err = transit_state_mb (preg, state, mctx);
+ *err = transit_state_mb (dfa, state, mctx);
if (BE (*err != REG_NOERROR, 0))
return NULL;
}
@@ -2160,7 +2148,7 @@ transit_state (err, preg, mctx, state)
trtable = state->trtable;
if (trtable == NULL)
{
- trtable = build_trtable (preg, state);
+ trtable = build_trtable (dfa, state);
if (trtable == NULL)
{
*err = REG_ESPACE;
@@ -2186,7 +2174,7 @@ transit_state (err, preg, mctx, state)
else
{
/* don't use transition table */
- next_state = transit_state_sb (err, preg, state, mctx);
+ next_state = transit_state_sb (err, dfa, state, mctx);
if (BE (next_state == NULL && err != REG_NOERROR, 0))
return NULL;
}
@@ -2256,7 +2244,7 @@ transit_state (err, preg, mctx, state)
/* If the next state has back references. */
if (next_state->has_backref)
{
- *err = transit_state_bkref (preg, &next_state->nodes, mctx);
+ *err = transit_state_bkref (dfa, &next_state->nodes, mctx);
if (BE (*err != REG_NOERROR, 0))
return NULL;
next_state = mctx->state_log[cur_idx];
@@ -2307,13 +2295,12 @@ check_subexp_matching_top (dfa, mctx, cu
accepting the current input byte. */
static re_dfastate_t *
-transit_state_sb (err, preg, state, mctx)
+transit_state_sb (err, dfa, state, mctx)
reg_errcode_t *err;
- const regex_t *preg;
+ re_dfa_t *dfa;
re_dfastate_t *state;
re_match_context_t *mctx;
{
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
re_node_set next_nodes;
re_dfastate_t *next_state;
int node_cnt, cur_str_idx = re_string_cur_idx (&mctx->input);
@@ -2325,7 +2312,7 @@ transit_state_sb (err, preg, state, mctx
for (node_cnt = 0; node_cnt < state->nodes.nelem; ++node_cnt)
{
int cur_node = state->nodes.elems[node_cnt];
- if (check_node_accept (preg, dfa->nodes + cur_node, mctx, cur_str_idx))
+ if (check_node_accept (dfa, dfa->nodes + cur_node, mctx, cur_str_idx))
{
*err = re_node_set_merge (&next_nodes,
dfa->eclosures + dfa->nexts[cur_node]);
@@ -2349,13 +2336,12 @@ transit_state_sb (err, preg, state, mctx
#ifdef RE_ENABLE_I18N
static reg_errcode_t
-transit_state_mb (preg, pstate, mctx)
- const regex_t *preg;
+transit_state_mb (dfa, pstate, mctx)
+ re_dfa_t *dfa;
re_dfastate_t *pstate;
re_match_context_t *mctx;
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
int i;
for (i = 0; i < pstate->nodes.nelem; ++i)
@@ -2378,7 +2364,7 @@ transit_state_mb (preg, pstate, mctx)
/* How many bytes the node can accept? */
if (ACCEPT_MB_NODE (dfa->nodes[cur_node_idx].type))
- naccepted = check_node_accept_bytes (preg, cur_node_idx, &mctx->input,
+ naccepted = check_node_accept_bytes (dfa, cur_node_idx, &mctx->input,
re_string_cur_idx (&mctx->input));
if (naccepted == 0)
continue;
@@ -2420,13 +2406,12 @@ transit_state_mb (preg, pstate, mctx)
#endif /* RE_ENABLE_I18N */
static reg_errcode_t
-transit_state_bkref (preg, nodes, mctx)
- const regex_t *preg;
+transit_state_bkref (dfa, nodes, mctx)
+ re_dfa_t *dfa;
const re_node_set *nodes;
re_match_context_t *mctx;
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
int i;
int cur_str_idx = re_string_cur_idx (&mctx->input);
@@ -2453,7 +2438,7 @@ transit_state_bkref (preg, nodes, mctx)
/* `node' is a backreference.
Check the substring which the substring matched. */
bkc_idx = mctx->nbkref_ents;
- err = get_subexp (preg, mctx, node_idx, cur_str_idx);
+ err = get_subexp (dfa, mctx, node_idx, cur_str_idx);
if (BE (err != REG_NOERROR, 0))
goto free_return;
@@ -2518,7 +2503,7 @@ transit_state_bkref (preg, nodes, mctx)
cur_str_idx);
if (BE (err != REG_NOERROR, 0))
goto free_return;
- err = transit_state_bkref (preg, new_dest_nodes, mctx);
+ err = transit_state_bkref (dfa, new_dest_nodes, mctx);
if (BE (err != REG_NOERROR, 0))
goto free_return;
}
@@ -2536,13 +2521,12 @@ transit_state_bkref (preg, nodes, mctx)
delay these checking for prune_impossible_nodes(). */
static reg_errcode_t
-get_subexp (preg, mctx, bkref_node, bkref_str_idx)
- const regex_t *preg;
+get_subexp (dfa, mctx, bkref_node, bkref_str_idx)
+ re_dfa_t *dfa;
re_match_context_t *mctx;
int bkref_node, bkref_str_idx;
{
int subexp_num, sub_top_idx;
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
const char *buf = (const char *) re_string_get_buffer (&mctx->input);
/* Return if we have already checked BKREF_NODE at BKREF_STR_IDX. */
int cache_idx = search_cur_bkref_entry (mctx, bkref_str_idx);
@@ -2585,7 +2569,7 @@ get_subexp (preg, mctx, bkref_node, bkre
break; /* We don't need to search this sub expression any more. */
bkref_str_off += sl_str_diff;
sl_str += sl_str_diff;
- err = get_subexp_sub (preg, mctx, sub_top, sub_last, bkref_node,
+ err = get_subexp_sub (dfa, mctx, sub_top, sub_last, bkref_node,
bkref_str_idx);
/* Reload buf, since the preceding call might have reallocated
@@ -2629,7 +2613,7 @@ get_subexp (preg, mctx, bkref_node, bkre
}
/* Can the OP_OPEN_SUBEXP node arrive the OP_CLOSE_SUBEXP node
in the current context? */
- err = check_arrival (preg, mctx, sub_top->path, sub_top->node,
+ err = check_arrival (dfa, mctx, sub_top->path, sub_top->node,
sub_top->str_idx, cls_node, sl_str, OP_CLOSE_SUBEXP);
if (err == REG_NOMATCH)
continue;
@@ -2638,7 +2622,7 @@ get_subexp (preg, mctx, bkref_node, bkre
sub_last = match_ctx_add_sublast (sub_top, cls_node, sl_str);
if (BE (sub_last == NULL, 0))
return REG_ESPACE;
- err = get_subexp_sub (preg, mctx, sub_top, sub_last, bkref_node,
+ err = get_subexp_sub (dfa, mctx, sub_top, sub_last, bkref_node,
bkref_str_idx);
if (err == REG_NOMATCH)
continue;
@@ -2654,8 +2638,8 @@ get_subexp (preg, mctx, bkref_node, bkre
and SUB_LAST. */
static reg_errcode_t
-get_subexp_sub (preg, mctx, sub_top, sub_last, bkref_node, bkref_str)
- const regex_t *preg;
+get_subexp_sub (dfa, mctx, sub_top, sub_last, bkref_node, bkref_str)
+ re_dfa_t *dfa;
re_match_context_t *mctx;
const re_sub_match_top_t *sub_top;
re_sub_match_last_t *sub_last;
@@ -2664,7 +2648,7 @@ get_subexp_sub (preg, mctx, sub_top, sub
reg_errcode_t err;
int to_idx;
/* Can the subexpression arrive the back reference? */
- err = check_arrival (preg, mctx, &sub_last->path, sub_last->node,
+ err = check_arrival (dfa, mctx, &sub_last->path, sub_last->node,
sub_last->str_idx, bkref_node, bkref_str, OP_OPEN_SUBEXP);
if (err != REG_NOERROR)
return err;
@@ -2709,14 +2693,13 @@ find_subexp_node (dfa, nodes, subexp_idx
Return REG_NOERROR if it can arrive, or REG_NOMATCH otherwise. */
static reg_errcode_t
-check_arrival (preg, mctx, path, top_node, top_str, last_node, last_str,
+check_arrival (dfa, mctx, path, top_node, top_str, last_node, last_str,
type)
- const regex_t *preg;
+ re_dfa_t *dfa;
re_match_context_t *mctx;
state_array_t *path;
int top_node, top_str, last_node, last_str, type;
{
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
reg_errcode_t err;
int subexp_num, backup_cur_idx, str_idx, null_cnt;
re_dfastate_t *cur_state = NULL;
@@ -2780,7 +2763,7 @@ check_arrival (preg, mctx, path, top_nod
{
if (next_nodes.nelem)
{
- err = expand_bkref_cache (preg, mctx, &next_nodes, str_idx, last_str,
+ err = expand_bkref_cache (dfa, mctx, &next_nodes, str_idx, last_str,
subexp_num, type);
if (BE ( err != REG_NOERROR, 0))
{
@@ -2812,7 +2795,7 @@ check_arrival (preg, mctx, path, top_nod
}
if (cur_state)
{
- err = check_arrival_add_next_nodes (preg, dfa, mctx, str_idx,
+ err = check_arrival_add_next_nodes (dfa, mctx, str_idx,
&cur_state->nodes, &next_nodes);
if (BE (err != REG_NOERROR, 0))
{
@@ -2829,7 +2812,7 @@ check_arrival (preg, mctx, path, top_nod
re_node_set_free (&next_nodes);
return err;
}
- err = expand_bkref_cache (preg, mctx, &next_nodes, str_idx, last_str,
+ err = expand_bkref_cache (dfa, mctx, &next_nodes, str_idx, last_str,
subexp_num, type);
if (BE ( err != REG_NOERROR, 0))
{
@@ -2872,8 +2855,7 @@ check_arrival (preg, mctx, path, top_nod
Can't we unify them? */
static reg_errcode_t
-check_arrival_add_next_nodes (preg, dfa, mctx, str_idx, cur_nodes, next_nodes)
- const regex_t *preg;
+check_arrival_add_next_nodes (dfa, mctx, str_idx, cur_nodes, next_nodes)
re_dfa_t *dfa;
re_match_context_t *mctx;
int str_idx;
@@ -2894,7 +2876,7 @@ check_arrival_add_next_nodes (preg, dfa,
/* If the node may accept `multi byte'. */
if (ACCEPT_MB_NODE (type))
{
- naccepted = check_node_accept_bytes (preg, cur_node, &mctx->input,
+ naccepted = check_node_accept_bytes (dfa, cur_node, &mctx->input,
str_idx);
if (naccepted > 1)
{
@@ -2930,7 +2912,7 @@ check_arrival_add_next_nodes (preg, dfa,
}
#endif /* RE_ENABLE_I18N */
if (naccepted
- || check_node_accept (preg, dfa->nodes + cur_node, mctx,
+ || check_node_accept (dfa, dfa->nodes + cur_node, mctx,
str_idx))
{
err = re_node_set_insert (next_nodes, dfa->nexts[cur_node]);
@@ -3051,15 +3033,14 @@ check_arrival_expand_ecl_sub (dfa, dst_n
in MCTX->BKREF_ENTS. */
static reg_errcode_t
-expand_bkref_cache (preg, mctx, cur_nodes, cur_str, last_str, subexp_num,
+expand_bkref_cache (dfa, mctx, cur_nodes, cur_str, last_str, subexp_num,
type)
- const regex_t *preg;
+ re_dfa_t *dfa;
re_match_context_t *mctx;
int cur_str, last_str, subexp_num, type;
re_node_set *cur_nodes;
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
int cache_idx, cache_idx_start;
/* The current state. */
@@ -3141,12 +3122,11 @@ expand_bkref_cache (preg, mctx, cur_node
Return the new table if succeeded, otherwise return NULL. */
static re_dfastate_t **
-build_trtable (preg, state)
- const regex_t *preg;
+build_trtable (dfa, state)
+ re_dfa_t *dfa;
re_dfastate_t *state;
{
reg_errcode_t err;
- re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
int i, j, ch;
unsigned int elem, mask;
int dests_node_malloced = 0, dest_states_malloced = 0;
@@ -3181,7 +3161,7 @@ build_trtable (preg, state)
/* At first, group all nodes belonging to `state' into several
destinations. */
- ndests = group_nodes_into_DFAstates (preg, state, dests_node, dests_ch);
+ ndests = group_nodes_into_DFAstates (dfa, state, dests_node, dests_ch);
if (BE (ndests <= 0, 0))
{
if (dests_node_malloced)
@@ -3369,14 +3349,13 @@ out_free:
to DEST_CH[i]. This function return the number of destinations. */
static int
-group_nodes_into_DFAstates (preg, state, dests_node, dests_ch)
- const regex_t *preg;
+group_nodes_into_DFAstates (dfa, state, dests_node, dests_ch)
+ re_dfa_t *dfa;
const re_dfastate_t *state;
re_node_set *dests_node;
bitset *dests_ch;
{
reg_errcode_t err;
- const re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
int i, j, k;
int ndests; /* Number of the destinations from `state'. */
bitset accepts; /* Characters a node can accept. */
@@ -3406,18 +3385,18 @@ group_nodes_into_DFAstates (preg, state,
else
#endif
bitset_set_all (accepts);
- if (!(preg->syntax & RE_DOT_NEWLINE))
+ if (!(dfa->syntax & RE_DOT_NEWLINE))
bitset_clear (accepts, '\n');
- if (preg->syntax & RE_DOT_NOT_NULL)
+ if (dfa->syntax & RE_DOT_NOT_NULL)
bitset_clear (accepts, '\0');
}
#ifdef RE_ENABLE_I18N
else if (type == OP_UTF8_PERIOD)
{
memset (accepts, 255, sizeof (unsigned int) * BITSET_UINTS / 2);
- if (!(preg->syntax & RE_DOT_NEWLINE))
+ if (!(dfa->syntax & RE_DOT_NEWLINE))
bitset_clear (accepts, '\n');
- if (preg->syntax & RE_DOT_NOT_NULL)
+ if (dfa->syntax & RE_DOT_NOT_NULL)
bitset_clear (accepts, '\0');
}
#endif
@@ -3561,12 +3540,11 @@ group_nodes_into_DFAstates (preg, state,
can only accept one byte. */
static int
-check_node_accept_bytes (preg, node_idx, input, str_idx)
- const regex_t *preg;
+check_node_accept_bytes (dfa, node_idx, input, str_idx)
+ re_dfa_t *dfa;
int node_idx, str_idx;
const re_string_t *input;
{
- const re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
const re_token_t *node = dfa->nodes + node_idx;
int char_len, elem_len;
int i;
@@ -3630,9 +3608,9 @@ check_node_accept_bytes (preg, node_idx,
/* FIXME: I don't think this if is needed, as both '\n'
and '\0' are char_len == 1. */
/* '.' accepts any one character except the following two cases. */
- if ((!(preg->syntax & RE_DOT_NEWLINE) &&
+ if ((!(dfa->syntax & RE_DOT_NEWLINE) &&
re_string_byte_at (input, str_idx) == '\n') ||
- ((preg->syntax & RE_DOT_NOT_NULL) &&
+ ((dfa->syntax & RE_DOT_NOT_NULL) &&
re_string_byte_at (input, str_idx) == '\0'))
return 0;
return char_len;
@@ -3860,8 +3838,8 @@ find_collation_sequence_value (mbs, mbs_
byte of the INPUT. */
static int
-check_node_accept (preg, node, mctx, idx)
- const regex_t *preg;
+check_node_accept (dfa, node, mctx, idx)
+ re_dfa_t *dfa;
const re_token_t *node;
const re_match_context_t *mctx;
int idx;
@@ -3890,8 +3868,8 @@ check_node_accept (preg, node, mctx, idx
/* FALLTHROUGH */
#endif
case OP_PERIOD:
- return !((ch == '\n' && !(preg->syntax & RE_DOT_NEWLINE))
- || (ch == '\0' && (preg->syntax & RE_DOT_NOT_NULL)));
+ return !((ch == '\n' && !(dfa->syntax & RE_DOT_NEWLINE))
+ || (ch == '\0' && (dfa->syntax & RE_DOT_NOT_NULL)));
default:
return 0;
}
Jakub
More information about the Libc-hacker
mailing list