SRE_NOAPI sre_regex_range_t *
sre_regex_turn_char_class_caseless(sre_pool_t *pool, sre_regex_range_t *range)
{
sre_char from, to;
sre_regex_range_t *r, *nr;
for (r = range; r; r = r->next) {
from = r->from;
to = r->to;
if (to >= 'A' && from <= 'Z') {
/* overlap with A-Z */
nr = sre_palloc(pool, sizeof(sre_regex_range_t));
if (nr == NULL) {
return NULL;
}
nr->from = sre_max(from, 'A') + 32;
nr->to = sre_min(to, 'Z') + 32;
nr->next = r->next;
r->next = nr;
r = nr;
}
if (to >= 'a' && from <= 'z') {
/* overlap with a-z */
nr = sre_palloc(pool, sizeof(sre_regex_range_t));
if (nr == NULL) {
return NULL;
}
nr->from = sre_max(from, 'a') - 32;
nr->to = sre_min(to, 'z') - 32;
nr->next = r->next;
r->next = nr;
r = nr;
}
}
return range;
}
SRE_API sre_vm_thompson_ctx_t *
sre_vm_thompson_jit_create_ctx(sre_pool_t *pool, sre_program_t *prog)
{
sre_uint_t size, len;
sre_vm_thompson_ctx_t *ctx;
sre_vm_thompson_thread_list_t *clist, *nlist;
size = sre_vm_thompson_jit_get_threads_added_size(prog);
dd("threads_added size: %d", (int) size);
ctx = sre_palloc(pool, sizeof(sre_vm_thompson_ctx_t) - 1 + size);
if (ctx == NULL) {
return NULL;
}
dd("ctx->threads_added: %x",
(int) offsetof(sre_vm_thompson_ctx_t, threads_added));
ctx->pool = pool;
ctx->program = prog;
len = prog->uniq_threads;
clist = sre_vm_thompson_create_thread_list(pool, len);
if (clist == NULL) {
return NULL;
}
ctx->current_threads = clist;
nlist = sre_vm_thompson_create_thread_list(pool, len);
if (nlist == NULL) {
return NULL;
}
ctx->next_threads = nlist;
ctx->tag = prog->tag + 1;
ctx->first_buf = 1;
return ctx;
}
sre_vm_thompson_ctx_t *
sre_vm_thompson_init(sre_pool_t *pool, sre_program_t *prog)
{
unsigned len;
sre_vm_thompson_ctx_t *ctx;
sre_vm_thompson_thread_list_t *clist, *nlist;
ctx = sre_palloc(pool, sizeof(sre_vm_thompson_ctx_t));
if (ctx == NULL) {
return NULL;
}
ctx->pool = pool;
ctx->program = prog;
len = prog->len;
clist = sre_vm_thompson_thread_list_create(pool, len);
if (clist == NULL) {
return NULL;
}
ctx->current_threads = clist;
nlist = sre_vm_thompson_thread_list_create(pool, len);
if (nlist == NULL) {
return NULL;
}
ctx->next_threads = nlist;
ctx->tag = prog->tag + 1;
ctx->first_buf = 1;
return ctx;
}
static sre_int_t
sre_program_get_leading_bytes_helper(sre_pool_t *pool, sre_instruction_t *pc,
sre_program_t *prog, sre_chain_t **res, unsigned tag)
{
sre_int_t rc;
sre_chain_t *cl, *ncl;
sre_instruction_t *bc;
if (pc->tag == tag) {
return SRE_OK;
}
if (pc == prog->start + 1) {
/* skip the dot (.) in the initial boilerplate ".*?" */
return SRE_OK;
}
pc->tag = tag;
switch (pc->opcode) {
case SRE_OPCODE_SPLIT:
rc = sre_program_get_leading_bytes_helper(pool, pc->x, prog, res,
tag);
if (rc != SRE_OK) {
return rc;
}
return sre_program_get_leading_bytes_helper(pool, pc->y, prog, res,
tag);
case SRE_OPCODE_JMP:
return sre_program_get_leading_bytes_helper(pool, pc->x, prog, res,
tag);
case SRE_OPCODE_SAVE:
if (++pc == prog->start + prog->len) {
return SRE_OK;
}
return sre_program_get_leading_bytes_helper(pool, pc, prog, res,
tag);
case SRE_OPCODE_MATCH:
prog->nullable = 1;
return SRE_DONE;
case SRE_OPCODE_ASSERT:
if (++pc == prog->start + prog->len) {
return SRE_OK;
}
return sre_program_get_leading_bytes_helper(pool, pc, prog, res, tag);
case SRE_OPCODE_ANY:
return SRE_DECLINED;
default:
/* CHAR, ANY, IN, NOTIN */
ncl = sre_palloc(pool, sizeof(sre_chain_t));
if (ncl == NULL) {
return SRE_ERROR;
}
ncl->data = pc;
ncl->next = NULL;
if (*res) {
for (cl = *res; /* void */; cl = cl->next) {
bc = cl->data;
if (bc->opcode == pc->opcode) {
if (bc->opcode == SRE_OPCODE_CHAR) {
if (bc->v.ch == pc->v.ch) {
return SRE_OK;
}
}
}
if (cl->next == NULL) {
cl->next = ncl;
return SRE_OK;
}
}
} else {
*res = ncl;
}
return SRE_OK;
}
/* impossible to reach here */
}
