static void acl_xml_node_reset(ACL_XML_NODE *node)
{
ACL_VSTRING_RESET(node->ltag);
ACL_VSTRING_RESET(node->rtag);
ACL_VSTRING_RESET(node->text);
ACL_VSTRING_TERMINATE(node->ltag);
ACL_VSTRING_TERMINATE(node->rtag);
ACL_VSTRING_TERMINATE(node->text);
node->id = NULL;
if (node->attr_list)
acl_array_clean(node->attr_list,
(void (*)(void*)) acl_xml_attr_free);
node->parent = NULL;
acl_ring_init(&node->children);
node->depth = 0;
acl_ring_init(&node->node);
node->curr_attr = NULL;
node->quote = 0;
node->last_ch = 0;
node->nlt = 0;
node->meta[0] = 0;
node->flag = 0;
node->status = ACL_XML_S_NXT;
}
static void incr_string(ACL_VSTRING *vp, int len, const char* s, int debug)
{
int i;
printf("max: %ld, len: %ld, cnt: %ld\r\n", (long) vp->maxlen,
vp->vbuf.len, vp->vbuf.cnt);
for (i = 0; i < len; i++)
ACL_VSTRING_ADDCH(vp, 'x');
if (s && *s)
acl_vstring_sprintf_append(vp, "%s", s);
else
ACL_VSTRING_TERMINATE(vp);
if (debug)
printf("[%s]\r\n", acl_vstring_str(vp));
printf("strlen: %ld, ACL_VSTRING_LEN: %ld, max: %ld\r\n",
(long) strlen(acl_vstring_str(vp)),
(long) ACL_VSTRING_LEN(vp), (long) vp->maxlen);
printf("Enter any key to continue ...\r\n\r\n");
getchar();
ACL_VSTRING_RESET(vp);
ACL_VSTRING_TERMINATE(vp);
}
static void __hdr_reset(HTTP_HDR_REQ *hh, int clear_cookies)
{
hh->port = 80;
hh->method[0] = 0;
hh->host[0] = 0;
hh->flag = 0;
if (hh->url_part) {
ACL_VSTRING_RESET(hh->url_part);
ACL_VSTRING_TERMINATE(hh->url_part);
}
if (hh->url_path) {
ACL_VSTRING_RESET(hh->url_path);
ACL_VSTRING_TERMINATE(hh->url_path);
}
if (hh->url_params) {
ACL_VSTRING_RESET(hh->url_params);
ACL_VSTRING_TERMINATE(hh->url_params);
}
if (hh->file_path) {
ACL_VSTRING_RESET(hh->file_path);
ACL_VSTRING_TERMINATE(hh->file_path);
}
if (hh->params_table)
acl_htable_reset(hh->params_table, __request_args_free_fn);
if (clear_cookies && hh->cookies_table)
acl_htable_reset(hh->cookies_table, __cookies_args_free_fn);
}
static const char *xml_parse_meta_text(ACL_XML *xml, const char *data)
{
int ch;
if (LEN(xml->curr_node->text) == 0) {
SKIP_SPACE(data);
}
while ((ch = *data) != 0) {
if (xml->curr_node->quote) {
if (ch == xml->curr_node->quote) {
xml->curr_node->quote = 0;
}
ACL_VSTRING_ADDCH(xml->curr_node->text, ch);
} else if (IS_QUOTE(ch)) {
if (xml->curr_node->quote == 0) {
xml->curr_node->quote = ch;
}
ACL_VSTRING_ADDCH(xml->curr_node->text, ch);
} else if (ch == '<') {
xml->curr_node->nlt++;
ACL_VSTRING_ADDCH(xml->curr_node->text, ch);
} else if (ch == '>') {
if (xml->curr_node->nlt == 0) {
char *last;
size_t off;
data++;
xml->curr_node->status = ACL_XML_S_MEND;
if ((xml->curr_node->flag & ACL_XML_F_META_QM) == 0)
break;
last = acl_vstring_end(xml->curr_node->text) - 1;
if (last < STR(xml->curr_node->text) || *last != '?')
break;
off = ACL_VSTRING_LEN(xml->curr_node->text) - 1;
if (off == 0)
break;
ACL_VSTRING_AT_OFFSET(xml->curr_node->text, off);
ACL_VSTRING_TERMINATE(xml->curr_node->text);
xml_meta_attr(xml->curr_node);
break;
}
xml->curr_node->nlt--;
ACL_VSTRING_ADDCH(xml->curr_node->text, ch);
} else {
ACL_VSTRING_ADDCH(xml->curr_node->text, ch);
}
data++;
}
ACL_VSTRING_TERMINATE(xml->curr_node->text);
return (data);
}
int acl_xml_decode(const char *in, ACL_VSTRING *out)
{
int n = 0, len;
const char *ptr = in, *pre;
const ACL_TOKEN *token;
const XML_SPEC *spec;
acl_pthread_once(&__token_once, xml_decode_init);
if (__token_tree == NULL)
acl_msg_fatal("__token_tree null");
while (*ptr != 0) {
pre = ptr;
token = acl_token_tree_match(__token_tree, &ptr, NULL, NULL);
if (token == NULL) {
pre = markup_unescape(pre, out);
len = (int) (ptr - pre);
if (len > 0)
acl_vstring_memcat(out, pre, len);
break;
}
spec = (const XML_SPEC*) token->ctx;
acl_assert(spec != NULL);
len = (int) (ptr - pre - spec->len);
if (len > 0)
acl_vstring_memcat(out, pre, len);
acl_vstring_strcat(out, spec->str);
n++;
}
ACL_VSTRING_TERMINATE(out);
return (n);
}
static const char *xml_parse_left_em(ACL_XML *xml, const char *data)
{
if (*data == '-') {
if (xml->curr_node->meta[1] != '-') {
xml->curr_node->meta[1] = '-';
} else if (xml->curr_node->meta[2] != '-') {
xml->curr_node->meta[0] = 0;
xml->curr_node->meta[1] = 0;
xml->curr_node->meta[2] = 0;
xml->curr_node->flag |= ACL_XML_F_META_CM;
xml->curr_node->status = ACL_XML_S_MCMT;
}
data++;
} else {
if (xml->curr_node->meta[1] == '-') {
ACL_VSTRING_ADDCH(xml->curr_node->ltag, '-');
xml->curr_node->meta[1] = 0;
}
xml->curr_node->flag |= ACL_XML_F_META_EM;
xml->curr_node->status = ACL_XML_S_MTAG;
}
ACL_VSTRING_TERMINATE(xml->curr_node->ltag);
return (data);
}
ACL_VSTRING *acl_hex_decode(ACL_VSTRING *result, const char *in, int len)
{
const unsigned char *cp;
int count;
unsigned int hex;
unsigned int bin;
ACL_VSTRING_RESET(result);
for (cp = UCHAR_PTR(in), count = len; count > 0; cp += 2, count -= 2) {
if (count < 2)
return (0);
hex = cp[0];
if (hex >= '0' && hex <= '9')
bin = (hex - '0') << 4;
else if (hex >= 'A' && hex <= 'F')
bin = (hex - 'A' + 10) << 4;
else if (hex >= 'a' && hex <= 'f')
bin = (hex - 'a' + 10) << 4;
else
return (0);
hex = cp[1];
if (hex >= '0' && hex <= '9')
bin |= (hex - '0') ;
else if (hex >= 'A' && hex <= 'F')
bin |= (hex - 'A' + 10) ;
else if (hex >= 'a' && hex <= 'f')
bin |= (hex - 'a' + 10) ;
else
return (0);
ACL_VSTRING_ADDCH(result, bin);
}
ACL_VSTRING_TERMINATE(result);
return (result);
}
void acl_token_name(const ACL_TOKEN *token, ACL_VSTRING *buf)
{
int i, n;
char *ptr, *pend, ch;
const ACL_TOKEN *token_iter;
ACL_VSTRING_RESET(buf);
token_iter = token;
while (token_iter && token_iter->parent != NULL) {
ACL_VSTRING_ADDCH(buf, token_iter->ch);
token_iter = token_iter->parent;
}
ACL_VSTRING_TERMINATE(buf);
pend = acl_vstring_end(buf) - 1;
ptr = STR(buf);
i = 0;
n = (int) (pend - ptr + 1) / 2;
while (i < n) {
ch = *ptr;
*ptr = *pend;
*pend = ch;
i++;
ptr++;
pend--;
}
}
static const char *tok822_comment(TOK822 *tp, const char *str)
{
int level = 1;
int ch;
/*
* XXX We cheat by storing comments in their external form. Otherwise it
* would be a royal pain to preserve \ before (. That would require a
* recursive parser; the easy to implement stack-based recursion would be
* too expensive.
*/
ACL_VSTRING_ADDCH(tp->vstr, '(');
while ((ch = *(const unsigned char *) str) != 0) {
ACL_VSTRING_ADDCH(tp->vstr, ch);
str++;
if (ch == '(') { /* comments can nest! */
level++;
} else if (ch == ')') {
if (--level == 0)
break;
} else if (ch == '\\') {
if ((ch = *(unsigned char *) str) == 0)
break;
ACL_VSTRING_ADDCH(tp->vstr, ch);
str++;
}
}
ACL_VSTRING_TERMINATE(tp->vstr);
return (str);
}
static const char *xml_parse_left_tag(ACL_XML *xml, const char *data)
{
int ch;
if (LEN(xml->curr_node->ltag) == 0) {
SKIP_SPACE(data);
}
while ((ch = *data) != 0) {
data++;
if (ch == '>') {
xml->curr_node->status = ACL_XML_S_LGT;
xml_parse_check_self_closed(xml);
if ((xml->curr_node->flag & ACL_XML_F_SELF_CL)
&& xml->curr_node->last_ch == '/')
{
acl_vstring_truncate(xml->curr_node->ltag,
LEN(xml->curr_node->ltag) - 1);
}
break;
} else if (IS_SPACE(ch)) {
xml->curr_node->status = ACL_XML_S_ATTR;
xml->curr_node->last_ch = ch;
break;
} else {
ACL_VSTRING_ADDCH(xml->curr_node->ltag, ch);
xml->curr_node->last_ch = ch;
}
}
ACL_VSTRING_TERMINATE(xml->curr_node->ltag);
return (data);
}
static const char *xml_meta_attr_value(ACL_XML_ATTR *attr, const char *data)
{
int ch;
SKIP_SPACE(data);
if (*data == 0)
return data;
if (IS_QUOTE(*data))
attr->quote = *data++;
while ((ch = *data) != 0) {
if (attr->backslash) {
if (ch == 'b')
ADDCH(attr->value, '\b');
else if (ch == 'f')
ADDCH(attr->value, '\f');
else if (ch == 'n')
ADDCH(attr->value, '\n');
else if (ch == 'r')
ADDCH(attr->value, '\r');
else if (ch == 't')
ADDCH(attr->value, '\t');
else
ADDCH(attr->value, ch);
attr->backslash = 0;
} else if (ch == '\\') {
if (attr->part_word) {
ADDCH(attr->value, ch);
attr->part_word = 0;
} else
attr->backslash = 1;
} else if (attr->quote) {
if (ch == attr->quote) {
data++;
break;
}
ADDCH(attr->value, ch);
} else if (IS_SPACE(ch)) {
data++;
break;
} else {
ADDCH(attr->value, ch);
if ((attr->node->xml->flag & ACL_XML_FLAG_PART_WORD)) {
if (attr->part_word)
attr->part_word = 0;
else if (ch < 0)
attr->part_word = 1;
}
}
data++;
}
ACL_VSTRING_TERMINATE(attr->value);
return data;
}
ACL_VSTRING *acl_hex_encode(ACL_VSTRING *result, const char *in, int len)
{
const unsigned char *cp;
int ch;
int count;
ACL_VSTRING_RESET(result);
for (cp = UCHAR_PTR(in), count = len; count > 0; count--, cp++) {
ch = *cp;
ACL_VSTRING_ADDCH(result, acl_hex_chars[(ch >> 4) & 0xf]);
ACL_VSTRING_ADDCH(result, acl_hex_chars[ch & 0xf]);
}
ACL_VSTRING_TERMINATE(result);
return (result);
}
static const char *xml_parse_meta_tag(ACL_XML *xml, const char *data)
{
int ch;
while ((ch = *data) != 0) {
data++;
if (IS_SPACE(ch)) {
xml->curr_node->status = ACL_XML_S_MTXT;
break;
}
ACL_VSTRING_ADDCH(xml->curr_node->ltag, ch);
}
ACL_VSTRING_TERMINATE(xml->curr_node->ltag);
return (data);
}
static const char *xml_parse_attr(ACL_XML *xml, const char *data)
{
int ch;
ACL_XML_ATTR *attr = xml->curr_node->curr_attr;
if (attr == NULL || LEN(attr->name) == 0) {
SKIP_SPACE(data); /* 略过 ' ', '\t' */
if (*data == 0)
return (NULL);
SKIP_WHILE(*data == '=', data);
if (*data == 0)
return (NULL);
}
if (*data == '>') {
xml->curr_node->status = ACL_XML_S_LGT;
xml_parse_check_self_closed(xml);
xml->curr_node->curr_attr = NULL;
data++;
return (data);
}
xml->curr_node->last_ch = *data;
if (*data == '/') {
data++;
return (data);
}
if (attr == NULL) {
attr = acl_xml_attr_alloc(xml->curr_node);
xml->curr_node->curr_attr = attr;
}
while ((ch = *data) != 0) {
xml->curr_node->last_ch = ch;
if (ch == '=') {
xml->curr_node->status = ACL_XML_S_AVAL;
data++;
break;
}
if (!IS_SPACE(ch))
ACL_VSTRING_ADDCH(attr->name, ch);
data++;
}
ACL_VSTRING_TERMINATE(attr->name);
return (data);
}
void mime_state_downgrade(MIME_STATE *state, int rec_type,
const char *text, ssize_t len)
{
static char hexchars[] = "0123456789ABCDEF";
const unsigned char *cp;
MIME_NODE *node = state->curr_node;
int ch;
#define CU_CHAR_PTR(x) ((const unsigned char *) (x))
#define QP_ENCODE(buffer, ch) { \
buffer += '='; \
buffer += (char) hexchars[(ch >> 4) & 0xff]; \
buffer += (char) hexchars[ch & 0xf]; \
}
/*
* Insert a soft line break when the output reaches a critical length
* before we reach a hard line break.
*/
for (cp = CU_CHAR_PTR(text); cp < CU_CHAR_PTR(text + len); cp++) {
/* Critical length before hard line break. */
if (LEN(node->buffer) > 72) {
node->buffer += '=';
}
/* Append the next character. */
ch = *cp;
if ((ch < 32 && ch != '\t') || ch == '=' || ch > 126) {
QP_ENCODE(node->buffer, ch);
} else {
ADDCH(node->buffer, ch);
}
}
/*
* Flush output after a hard line break (i.e. the end of a REC_TYPE_NORM
* record). Fix trailing whitespace as per the RFC: in the worst case,
* the output length will grow from 73 characters to 75 characters.
*/
if (rec_type == REC_TYPE_NORM) {
if (LEN(node->buffer) > 0
&& ((ch = END(node->buffer)[-1]) == ' ' || ch == '\t'))
{
acl_vstring_truncate(node->buffer, LEN(node->buffer) - 1);
QP_ENCODE(node->buffer, ch);
}
ACL_VSTRING_TERMINATE(node->buffer);
}
}
static const char *xml_meta_attr_name(ACL_XML_ATTR *attr, const char *data)
{
int ch;
while ((ch = *data) != 0) {
if (ch == '=') {
data++;
ACL_VSTRING_TERMINATE(attr->name);
break;
}
if (!IS_SPACE(ch))
ACL_VSTRING_ADDCH(attr->name, ch);
data++;
}
return data;
}
int acl_xml_encode(const char *in, ACL_VSTRING *out)
{
int n = 0;
while (*in) {
if (__charmap[(unsigned char)(*in)] != NULL) {
acl_vstring_strcat(out,
__charmap[(unsigned char)(*in)]);
n++;
} else
ACL_VSTRING_ADDCH(out, (*in));
in++;
}
ACL_VSTRING_TERMINATE(out);
return (n);
}
ACL_VSTRING *escape(ACL_VSTRING *result, const char *data, ssize_t len)
{
int ch;
ACL_VSTRING_RESET(result);
while (len-- > 0) {
ch = *UCHAR(data++);
if (ACL_ISASCII(ch)) {
if (ACL_ISPRINT(ch)) {
if (ch == '\\')
ACL_VSTRING_ADDCH(result, ch);
ACL_VSTRING_ADDCH(result, ch);
continue;
} else if (ch == '\a') { /* \a -> audible bell */
acl_vstring_strcat(result, "\\a");
continue;
} else if (ch == '\b') { /* \b -> backspace */
acl_vstring_strcat(result, "\\b");
continue;
} else if (ch == '\f') { /* \f -> formfeed */
acl_vstring_strcat(result, "\\f");
continue;
} else if (ch == '\n') { /* \n -> newline */
acl_vstring_strcat(result, "\\n");
continue;
} else if (ch == '\r') { /* \r -> carriagereturn */
acl_vstring_strcat(result, "\\r");
continue;
} else if (ch == '\t') { /* \t -> horizontal tab */
acl_vstring_strcat(result, "\\t");
continue;
} else if (ch == '\v') { /* \v -> vertical tab */
acl_vstring_strcat(result, "\\v");
continue;
}
}
if (ACL_ISDIGIT(*UCHAR(data)))
acl_vstring_sprintf_append(result, "\\%03d", ch);
else
acl_vstring_sprintf_append(result, "\\%d", ch);
}
ACL_VSTRING_TERMINATE(result);
return (result);
}
ACL_VSTRING *tok822_internalize(ACL_VSTRING *vp, TOK822 *tree, int flags)
{
TOK822 *tp;
if (flags & TOK822_STR_WIPE)
ACL_VSTRING_RESET(vp);
for (tp = tree; tp; tp = tp->next) {
switch (tp->type) {
case ',':
ACL_VSTRING_ADDCH(vp, tp->type);
if (flags & TOK822_STR_LINE) {
ACL_VSTRING_ADDCH(vp, '\n');
continue;
}
break;
case TOK822_ADDR:
tok822_internalize(vp, tp->head, TOK822_STR_NONE);
break;
case TOK822_COMMENT:
case TOK822_ATOM:
case TOK822_QSTRING:
acl_vstring_strcat(vp, acl_vstring_str(tp->vstr));
break;
case TOK822_DOMLIT:
ACL_VSTRING_ADDCH(vp, '[');
acl_vstring_strcat(vp, acl_vstring_str(tp->vstr));
ACL_VSTRING_ADDCH(vp, ']');
break;
case TOK822_STARTGRP:
ACL_VSTRING_ADDCH(vp, ':');
break;
default:
if (tp->type >= TOK822_MINTOK)
acl_msg_panic("tok822_internalize: unknown operator %d", tp->type);
ACL_VSTRING_ADDCH(vp, tp->type);
}
if (tok822_append_space(tp))
ACL_VSTRING_ADDCH(vp, ' ');
}
if (flags & TOK822_STR_TERM)
ACL_VSTRING_TERMINATE(vp);
return (vp);
}
const char *str_name_mask_opt(ACL_VSTRING *buf, const char *context,
const NAME_MASK *table,
int mask, int flags)
{
const char *myname = "name_mask";
const NAME_MASK *np;
int len;
static ACL_VSTRING *my_buf = 0;
int delim = (flags & NAME_MASK_COMMA ? ',' :
(flags & NAME_MASK_PIPE ? '|' : ' '));
if (buf == 0) {
if (my_buf == 0)
my_buf = acl_vstring_alloc(1);
buf = my_buf;
}
ACL_VSTRING_RESET(buf);
for (np = table; mask != 0; np++) {
if (np->name == 0) {
if (flags & NAME_MASK_FATAL) {
acl_msg_fatal("%s: unknown %s bit in mask: 0x%x",
myname, context, mask);
} else if (flags & NAME_MASK_RETURN) {
acl_msg_warn("%s: unknown %s bit in mask: 0x%x",
myname, context, mask);
return (0);
} else if (flags & NAME_MASK_NUMBER) {
acl_vstring_sprintf_append(buf, "0x%x%c", mask, delim);
}
break;
}
if (mask & np->mask) {
mask &= ~np->mask;
acl_vstring_sprintf_append(buf, "%s%c", np->name, delim);
}
}
if ((len = (int) ACL_VSTRING_LEN(buf)) > 0)
acl_vstring_truncate(buf, len - 1);
ACL_VSTRING_TERMINATE(buf);
return (STR(buf));
}
static const char *xml_parse_text(ACL_XML *xml, const char *data)
{
int ch;
if (LEN(xml->curr_node->text) == 0) {
SKIP_SPACE(data);
if (*data == 0)
return (NULL);
}
while ((ch = *data) != 0) {
if (ch == '<') {
xml->curr_node->status = ACL_XML_S_RLT;
data++;
break;
}
ACL_VSTRING_ADDCH(xml->curr_node->text, ch);
data++;
}
ACL_VSTRING_TERMINATE(xml->curr_node->text);
if (xml->curr_node->status != ACL_XML_S_RLT)
return (data);
if ((xml->curr_node->flag & ACL_XML_F_SELF_CL)) {
/* 如果该标签是自关闭类型，则应使父结点直接跳至右边 '/' 处理位置,
* 同时使本结点跳至右边 '>' 处理位置
*/
ACL_XML_NODE *parent = acl_xml_node_parent(xml->curr_node);
if (parent != xml->root)
parent->status = ACL_XML_S_RLT;
xml->curr_node->status = ACL_XML_S_RGT;
}
if (LEN(xml->curr_node->text) == 0)
return (data);
return (data);
}
static void strip_address(ACL_VSTRING *vp, ssize_t start, TOK822 *addr)
{
ACL_VSTRING *tmp;
/*
* Emit plain <address>. Discard any comments or phrases.
*/
ACL_VSTRING_TERMINATE(vp);
acl_msg_warn("stripping too many comments from address: %.100s...",
acl_vstring_str(vp) + start);
//acl_printable(vstring_str(vp) + start, '?')); //zsx
acl_vstring_truncate(vp, start);
ACL_VSTRING_ADDCH(vp, '<');
if (addr) {
tmp = acl_vstring_alloc(100);
tok822_internalize(tmp, addr, TOK822_STR_TERM);
quote_822_local_flags(vp, acl_vstring_str(tmp),
QUOTE_FLAG_8BITCLEAN | QUOTE_FLAG_APPEND);
acl_vstring_free(tmp);
}
ACL_VSTRING_ADDCH(vp, '>');
}
int mac_expand(ACL_VSTRING *result, const char *pattern, int flags,
const char *filter,
MAC_EXP_LOOKUP_FN lookup, char *context)
{
MAC_EXP mc;
int status;
/*
* Bundle up the request and do the substitutions.
*/
mc.result = result;
mc.flags = flags;
mc.filter = filter;
mc.lookup = lookup;
mc.context = context;
mc.status = 0;
mc.level = 0;
ACL_VSTRING_RESET(result);
status = mac_parse(pattern, mac_expand_callback, (char *) &mc);
ACL_VSTRING_TERMINATE(result);
return (status);
}
int acl_xml_encode(const char *in, ACL_VSTRING *out)
{
const unsigned char *ptr = (const unsigned char*) in;
int n = 0;
size_t len = strlen(in);
len += len / 2;
ACL_VSTRING_SPACE(out, (int) len);
while (*ptr) {
if (__charmap[*ptr] != NULL) {
acl_vstring_strcat(out, __charmap[*ptr]);
n++;
} else {
ACL_VSTRING_ADDCH(out, *ptr);
}
ptr++;
}
ACL_VSTRING_TERMINATE(out);
return n;
}
static const char *xml_parse_right_lt(ACL_XML *xml, const char *data)
{
ACL_XML_NODE *node;
SKIP_SPACE(data);
if (*data == 0)
return (NULL);
if (*data == '/') {
xml->curr_node->status = ACL_XML_S_RTAG;
data++;
return (data);
} else if ((xml->curr_node->flag & ACL_XML_F_LEAF)) {
ACL_VSTRING_ADDCH(xml->curr_node->text, '<');
ACL_VSTRING_ADDCH(xml->curr_node->text, *data);
ACL_VSTRING_TERMINATE(xml->curr_node->text);
xml->curr_node->status = ACL_XML_S_TXT;
data++;
return (data);
}
/* 说明遇到了当前结点的子结点 */
/* 重新设置当前结点状态，以便于其可以找到 "</" */
xml->curr_node->status = ACL_XML_S_TXT;
/* 创建新的子结点，并将其加入至当前结点的子结点集合中 */
node = acl_xml_node_alloc(xml);
acl_xml_node_add_child(xml->curr_node, node);
node->depth = xml->curr_node->depth + 1;
if (node->depth > xml->depth)
xml->depth = node->depth;
xml->curr_node = node;
xml->curr_node->status = ACL_XML_S_LLT;
return (data);
}
HTTP_HDR_REQ *http_hdr_req_create(const char *url,
const char *method, const char *version)
{
const char *myname = "http_hdr_req_create";
HTTP_HDR_REQ *hdr_req;
ACL_VSTRING *req_line = acl_vstring_alloc(256);
HTTP_HDR_ENTRY *entry;
const char *ptr;
static char *__user_agent = "Mozilla/5.0 (Windows; U; Windows NT 5.0"
"; zh-CN; rv:1.9.0.3) Gecko/2008092417 ACL/3.0.6";
if (url == NULL || *url == 0) {
acl_msg_error("%s(%d): url invalid", myname, __LINE__);
return NULL;
}
if (method == NULL || *method == 0) {
acl_msg_error("%s(%d): method invalid", myname, __LINE__);
return NULL;
}
if (version == NULL || *version == 0) {
acl_msg_error("%s(%d): version invalid", myname, __LINE__);
return NULL;
}
acl_vstring_strcpy(req_line, method);
acl_vstring_strcat(req_line, " ");
if (strncasecmp(url, "http://", sizeof("http://") - 1) == 0)
url += sizeof("http://") - 1;
else if (strncasecmp(url, "https://", sizeof("https://") - 1) == 0)
url += sizeof("https://") -1;
ptr = strchr(url, '/');
if (ptr)
acl_vstring_strcat(req_line, ptr);
else {
ACL_VSTRING_ADDCH(req_line, '/');
ACL_VSTRING_TERMINATE(req_line);
}
acl_vstring_strcat(req_line, " ");
acl_vstring_strcat(req_line, version);
entry = http_hdr_entry_new(acl_vstring_str(req_line));
acl_vstring_free(req_line);
if (entry == NULL) {
acl_msg_error("%s(%d): http_hdr_entry_new return null for (%s)",
myname, __LINE__, acl_vstring_str(req_line));
return NULL;
}
hdr_req = http_hdr_req_new();
http_hdr_append_entry(&hdr_req->hdr, entry);
hdr_req->flag |= (HTTP_HDR_REQ_FLAG_PARSE_PARAMS | HTTP_HDR_REQ_FLAG_PARSE_COOKIE);
if (http_hdr_req_line_parse(hdr_req) < 0) {
http_hdr_req_free(hdr_req);
return NULL;
}
hdr_req->host[0] = 0;
__get_host_from_url(hdr_req->host, sizeof(hdr_req->host), url);
if (hdr_req->host[0] != 0)
http_hdr_put_str(&hdr_req->hdr, "Host", hdr_req->host);
http_hdr_put_str(&hdr_req->hdr, "Connection", "Close");
http_hdr_put_str(&hdr_req->hdr, "User-Agent", __user_agent);
return hdr_req;
}
static int vstring_extend(ACL_VBUF *bp, ssize_t incr)
{
const char *myname = "vstring_extend";
ssize_t used = (ssize_t) (bp->ptr - bp->data), new_len;
ACL_VSTRING *vp = (ACL_VSTRING *) bp;
if (vp->maxlen > 0 && (ssize_t) ACL_VSTRING_LEN(vp) >= vp->maxlen) {
ACL_VSTRING_AT_OFFSET(vp, vp->maxlen - 1);
ACL_VSTRING_TERMINATE(vp);
acl_msg_warn("%s(%d), %s: overflow maxlen: %ld, %ld",
__FILE__, __LINE__, myname, (long) vp->maxlen,
(long) ACL_VSTRING_LEN(vp));
bp->flags |= ACL_VBUF_FLAG_EOF;
return ACL_VBUF_EOF;
}
#ifdef ACL_WINDOWS
if (bp->fd == ACL_FILE_INVALID && (bp->flags & ACL_VBUF_FLAG_FIXED))
#else
if (bp->fd < 0 && (bp->flags & ACL_VBUF_FLAG_FIXED))
#endif
{
acl_msg_warn("%s(%d), %s: can't extend fixed buffer",
__FILE__, __LINE__, myname);
return ACL_VBUF_EOF;
}
/*
* Note: vp->vbuf.len is the current buffer size (both on entry and on
* exit of this routine). We round up the increment size to the buffer
* size to avoid silly little buffer increments. With really large
* strings we might want to abandon the length doubling strategy, and
* go to fixed increments.
*/
#ifdef INCR_NO_DOUBLE
/* below come from redis-server/sds.c/sdsMakeRoomFor, which can
* avoid memory double growing too large --- 2015.2.2, zsx
*/
new_len = bp->len + incr;
if (new_len < MAX_PREALLOC)
new_len *= 2;
else
new_len += MAX_PREALLOC;
#else
new_len = bp->len + (bp->len > incr ? bp->len : incr);
#endif
if (vp->maxlen > 0 && new_len > vp->maxlen)
new_len = vp->maxlen;
if (vp->slice)
bp->data = (unsigned char *) acl_slice_pool_realloc(
__FILE__, __LINE__, vp->slice, bp->data, new_len);
else if (vp->dbuf) {
const unsigned char *data = bp->data;
bp->data = (unsigned char *) acl_dbuf_pool_alloc(
vp->dbuf, new_len);
memcpy(bp->data, data, used);
acl_dbuf_pool_free(vp->dbuf, data);
} else if (bp->fd != ACL_FILE_INVALID) {
#ifdef ACL_UNIX
acl_off_t off = new_len - 1;
if (acl_lseek(bp->fd, off, SEEK_SET) != (acl_off_t) off)
acl_msg_fatal("lseek failed: %s, off: %lld",
acl_last_serror(), off);
if (acl_file_write(bp->fd, "\0", 1, 0, NULL, NULL)
== ACL_VSTREAM_EOF)
{
acl_msg_fatal("write error: %s", acl_last_serror());
}
#endif
} else
bp->data = (unsigned char *) acl_myrealloc(bp->data, new_len);
bp->len = new_len;
bp->ptr = bp->data + used;
bp->cnt = bp->len - used;
return 0;
}
ACL_VSTRING *tok822_externalize(ACL_VSTRING *vp, TOK822 *tree, int flags)
{
ACL_VSTRING *tmp;
TOK822 *tp;
ssize_t start = 0;
TOK822 *addr = 0;
ssize_t addr_len = 0;
/*
* Guard against a Sendmail buffer overflow (CERT advisory CA-2003-07).
* The problem was that Sendmail could store too much non-address text
* (comments, phrases, etc.) into a static 256-byte buffer.
*
* When the buffer fills up, fixed Sendmail versions remove comments etc.
* and reduce the information to just <$g>, which expands to <address>.
* No change is made when an address expression (text separated by
* commas) contains no address. This fix reportedly also protects
* Sendmail systems that are still vulnerable to this problem.
*
* Postfix takes the same approach, grudgingly. To avoid unnecessary damage,
* Postfix removes comments etc. only when the amount of non-address text
* in an address expression (text separated by commas) exceeds 250 bytes.
*
* With Sendmail, the address part of an address expression is the
* right-most <> instance in that expression. If an address expression
* contains no <>, then Postfix guarantees that it contains at most one
* non-comment string; that string is the address part of the address
* expression, so there is no ambiguity.
*
* Finally, we note that stress testing shows that other code in Sendmail
* 8.12.8 bluntly truncates ``text <address>'' to 256 bytes even when
* this means chopping the <address> somewhere in the middle. This is a
* loss of control that we're not entirely comfortable with. However,
* unbalanced quotes and dangling backslash do not seem to influence the
* way that Sendmail parses headers, so this is not an urgent problem.
*/
#define MAX_NONADDR_LENGTH 250
#define RESET_NONADDR_LENGTH { \
start = ACL_VSTRING_LEN(vp); \
addr = 0; \
addr_len = 0; \
}
#define ENFORCE_NONADDR_LENGTH do { \
if (addr && (ssize_t) ACL_VSTRING_LEN(vp) - addr_len > start + MAX_NONADDR_LENGTH) \
strip_address(vp, start, addr->head); \
} while(0)
if (flags & TOK822_STR_WIPE)
ACL_VSTRING_RESET(vp);
if (flags & TOK822_STR_TRNC)
RESET_NONADDR_LENGTH;
for (tp = tree; tp; tp = tp->next) {
switch (tp->type) {
case ',':
if (flags & TOK822_STR_TRNC)
ENFORCE_NONADDR_LENGTH;
ACL_VSTRING_ADDCH(vp, tp->type);
ACL_VSTRING_ADDCH(vp, (flags & TOK822_STR_LINE) ? '\n' : ' ');
if (flags & TOK822_STR_TRNC)
RESET_NONADDR_LENGTH;
continue;
/*
* XXX In order to correctly externalize an address, it is not
* sufficient to quote individual atoms. There are higher-level
* rules that say when an address localpart needs to be quoted.
* We wing it with the quote_822_local() routine, which ignores
* the issue of atoms in the domain part that would need quoting.
*/
case TOK822_ADDR:
addr = tp;
tmp = acl_vstring_alloc(100);
tok822_internalize(tmp, tp->head, TOK822_STR_TERM);
addr_len = ACL_VSTRING_LEN(vp);
quote_822_local_flags(vp, acl_vstring_str(tmp),
QUOTE_FLAG_8BITCLEAN | QUOTE_FLAG_APPEND);
addr_len = ACL_VSTRING_LEN(vp) - addr_len;
acl_vstring_free(tmp);
break;
case TOK822_ATOM:
case TOK822_COMMENT:
acl_vstring_strcat(vp, acl_vstring_str(tp->vstr));
break;
case TOK822_QSTRING:
ACL_VSTRING_ADDCH(vp, '"');
tok822_copy_quoted(vp, acl_vstring_str(tp->vstr), "\"\\\r\n");
ACL_VSTRING_ADDCH(vp, '"');
break;
case TOK822_DOMLIT:
ACL_VSTRING_ADDCH(vp, '[');
tok822_copy_quoted(vp, acl_vstring_str(tp->vstr), "\\\r\n");
ACL_VSTRING_ADDCH(vp, ']');
break;
case TOK822_STARTGRP:
ACL_VSTRING_ADDCH(vp, ':');
break;
case '<':
if (tp->next && tp->next->type == '>') {
addr = tp;
addr_len = 0;
}
ACL_VSTRING_ADDCH(vp, '<');
break;
default:
if (tp->type >= TOK822_MINTOK)
acl_msg_panic("tok822_externalize: unknown operator %d", tp->type);
ACL_VSTRING_ADDCH(vp, tp->type);
}
if (tok822_append_space(tp))
ACL_VSTRING_ADDCH(vp, ' ');
}
if (flags & TOK822_STR_TRNC)
ENFORCE_NONADDR_LENGTH;
if (flags & TOK822_STR_TERM)
ACL_VSTRING_TERMINATE(vp);
return (vp);
}
static const char *json_tag(ACL_JSON *json, const char *data)
{
ACL_JSON_NODE *node = json->curr_node;
int ch;
while ((ch = *data) != 0) {
/* 如果前面有引号，则需要找到结尾引号 */
if (node->quote) {
if (node->backslash) {
if (ch == 'b')
ADDCH(node->ltag, '\b');
else if (ch == 'f')
ADDCH(node->ltag, '\f');
else if (ch == 'n')
ADDCH(node->ltag, '\n');
else if (ch == 'r')
ADDCH(node->ltag, '\r');
else if (ch == 't')
ADDCH(node->ltag, '\t');
else
ADDCH(node->ltag, ch);
node->backslash = 0;
}
/* 当为双字节汉字时，第一个字节为的高位为 1，
* 第二个字节为 92，正好与转义字符相同
*/
else if (ch == '\\') {
/* 处理半个汉字的情形 */
if (node->part_word) {
ADDCH(node->ltag, ch);
node->part_word = 0;
} else
node->backslash = 1;
} else if (ch == node->quote) {
ACL_JSON_NODE *parent;
parent = acl_json_node_parent(node);
acl_assert(parent);
/* 数组对象的子结点允许为单独的字符串或对象 */
if (parent->left_ch == '[')
json->status = ACL_JSON_S_NEXT;
/* 标签值分析结束，下一步需要找到冒号 */
else
json->status = ACL_JSON_S_COLON;
node->quote = 0;
node->part_word = 0;
data++;
break;
}
/* 是否兼容后半个汉字为转义符 '\' 的情况 */
else if ((json->flag & ACL_JSON_FLAG_PART_WORD)) {
ADDCH(node->ltag, ch);
/* 处理半个汉字的情形 */
if (node->part_word)
node->part_word = 0;
else if (ch < 0)
node->part_word = 1;
} else {
ADDCH(node->ltag, ch);
}
}
/* 分析标签名前没有引号的情况 */
else if (node->backslash) {
ADDCH(node->ltag, ch);
node->backslash = 0;
}
/* 当为双字节汉字时，第一个字节为的高位为 1，
* 第二个字节为 92，正好与转义字符相同
*/
else if (ch == '\\') {
/* 处理半个汉字的情形 */
if (node->part_word) {
ADDCH(node->ltag, ch);
node->part_word = 0;
} else
node->backslash = 1;
} else if (IS_SPACE(ch) || ch == ':') {
/* 标签名分析结束，下一步需要找到冒号 */
json->status = ACL_JSON_S_COLON;
node->part_word = 0;
break;
}
/* 是否兼容后半个汉字为转义符 '\' 的情况 */
else if ((json->flag & ACL_JSON_FLAG_PART_WORD)) {
ADDCH(node->ltag, ch);
/* 处理半个汉字的情形 */
if (node->part_word)
node->part_word = 0;
else if (ch < 0)
node->part_word = 1;
} else {
ADDCH(node->ltag, ch);
}
data++;
}
/* 如果标签名非空，则需要保证以 0 结尾 */
if (LEN(node->ltag) > 0)
ACL_VSTRING_TERMINATE(node->ltag);
return data;
}
static const char *json_string(ACL_JSON *json, const char *data)
{
ACL_JSON_NODE *node = json->curr_node;
int ch;
/* 当文本长度为 0 时，可以认为还未遇到有效的字符 */
if (LEN(node->text) == 0) {
/* 先过滤开头没用的空格 */
SKIP_SPACE(data);
if (*data == 0)
return NULL;
}
/* 说明本结点是叶结点 */
while ((ch = *data) != 0) {
/* 如果开始有引号，则需要以该引号作为结尾符 */
if (node->quote) {
if (node->backslash) {
if (ch == 'b')
ADDCH(node->text, '\b');
else if (ch == 'f')
ADDCH(node->text, '\f');
else if (ch == 'n')
ADDCH(node->text, '\n');
else if (ch == 'r')
ADDCH(node->text, '\r');
else if (ch == 't')
ADDCH(node->text, '\t');
else
ADDCH(node->text, ch);
node->backslash = 0;
}
/* 当为双字节汉字时，第一个字节为的高位为 1，
* 第二个字节有可能为 92，正好与转义字符相同
*/
else if (ch == '\\') {
/* 处理半个汉字的情况，如果前一个字节是前
* 半个汉字，则当前的转义符当作后半个汉字
*/
if (node->part_word) {
ADDCH(node->text, ch);
node->part_word = 0;
} else
node->backslash = 1;
} else if (ch == node->quote) {
node->quote = 0;
/* 切换至查询该结点的兄弟结点的过程 */
json->status = ACL_JSON_S_STREND;
node->part_word = 0;
data++;
break;
}
/* 是否兼容后半个汉字为转义符 '\' 的情况 */
else if ((json->flag & ACL_JSON_FLAG_PART_WORD)) {
ADDCH(node->text, ch);
/* 若前一个字节为前半个汉字，则当前字节
* 为后半个汉字，正好为一个完整的汉字
*/
if (node->part_word)
node->part_word = 0;
/* 前一个字节非前半个汉字且当前字节高位
* 为 1，则表明当前字节为前半个汉字
*/
else if (ch < 0)
node->part_word = 1;
} else {
ADDCH(node->text, ch);
}
} else if (node->backslash) {
ADDCH(node->text, ch);
node->backslash = 0;
} else if (ch == '\\') {
if (node->part_word) {
ADDCH(node->text, ch);
node->part_word = 0;
} else
node->backslash = 1;
} else if (IS_SPACE(ch) || ch == ',' || ch == ';'
|| ch == '}' || ch == ']')
{
/* 切换至查询该结点的兄弟结点的过程 */
json->status = ACL_JSON_S_STREND;
break;
}
/* 是否兼容后半个汉字为转义符 '\' 的情况 */
else if ((json->flag & ACL_JSON_FLAG_PART_WORD)) {
ADDCH(node->text, ch);
/* 处理半个汉字的情形 */
if (node->part_word)
node->part_word = 0;
else if (ch < 0)
node->part_word = 1;
} else {
ADDCH(node->text, ch);
}
data++;
}
if (LEN(node->text) > 0)
ACL_VSTRING_TERMINATE(node->text);
return data;
}