tb_size_t tb_xml_reader_next(tb_xml_reader_ref_t reader)
{
// check
tb_xml_reader_impl_t* impl = (tb_xml_reader_impl_t*)reader;
tb_assert_and_check_return_val(impl && impl->rstream, TB_XML_READER_EVENT_NONE);
// reset event
impl->event = TB_XML_READER_EVENT_NONE;
// next
while (!impl->event)
{
// peek character
tb_char_t* pc = tb_null;
if (!tb_stream_need(impl->rstream, (tb_byte_t**)&pc, 1) || !pc) break;
// is element?
if (*pc == '<')
{
// parse element: <...>
tb_char_t const* element = tb_xml_reader_element_parse(impl);
tb_assert_and_check_break(element);
// is document begin: <?xml version="..." charset=".." ?>
tb_size_t size = tb_string_size(&impl->element);
if (size > 4 && !tb_strnicmp(element, "?xml", 4))
{
// update event
impl->event = TB_XML_READER_EVENT_DOCUMENT;
// update version & charset
tb_xml_node_ref_t attr = (tb_xml_node_ref_t)tb_xml_reader_attributes(reader);
for (; attr; attr = attr->next)
{
if (!tb_string_cstricmp(&attr->name, "version")) tb_string_strcpy(&impl->version, &attr->data);
if (!tb_string_cstricmp(&attr->name, "encoding")) tb_string_strcpy(&impl->charset, &attr->data);
}
// transform stream => utf-8
if (tb_string_cstricmp(&impl->charset, "utf-8") && tb_string_cstricmp(&impl->charset, "utf8"))
{
// charset
tb_size_t charset = TB_CHARSET_TYPE_UTF8;
if (!tb_string_cstricmp(&impl->charset, "gb2312") || !tb_string_cstricmp(&impl->charset, "gbk"))
charset = TB_CHARSET_TYPE_GB2312;
else tb_trace_e("the charset: %s is not supported", tb_string_cstr(&impl->charset));
// init transform stream
if (charset != TB_CHARSET_TYPE_UTF8)
{
#ifdef TB_CONFIG_MODULE_HAVE_CHARSET
// init the filter stream
if (!impl->fstream) impl->fstream = tb_stream_init_filter_from_charset(impl->istream, charset, TB_CHARSET_TYPE_UTF8);
else
{
// ctrl stream
if (!tb_stream_ctrl(impl->fstream, TB_STREAM_CTRL_FLTR_SET_STREAM, impl->istream)) break;
// the filter
tb_stream_filter_ref_t filter = tb_null;
if (!tb_stream_ctrl(impl->fstream, TB_STREAM_CTRL_FLTR_GET_FILTER, &filter)) break;
tb_assert_and_check_break(filter);
// ctrl filter
if (!tb_stream_filter_ctrl(filter, TB_STREAM_FILTER_CTRL_CHARSET_SET_FTYPE, charset)) break;
}
// open the filter stream
if (impl->fstream && tb_stream_open(impl->fstream))
impl->rstream = impl->fstream;
tb_string_cstrcpy(&impl->charset, "utf-8");
#else
// trace
tb_trace_e("unicode type is not supported, please enable charset module config if you want to use it!");
#endif
}
}
}
// is document type: <!DOCTYPE ... >
else if (size > 8 && !tb_strnicmp(element, "!DOCTYPE", 8))
{
// update event
impl->event = TB_XML_READER_EVENT_DOCUMENT_TYPE;
}
// is element end: </name>
else if (size > 1 && element[0] == '/')
{
// check
tb_check_break(impl->level);
// update event
impl->event = TB_XML_READER_EVENT_ELEMENT_END;
// leave
impl->level--;
}
// is comment: <!-- text -->
else if (size >= 3 && !tb_strncmp(element, "!--", 3))
{
// no comment end?
if (element[size - 2] != '-' || element[size - 1] != '-')
{
// patch '>'
tb_string_chrcat(&impl->element, '>');
// seek to comment end
tb_char_t ch = '\0';
tb_int_t n = 0;
while ((ch = tb_stream_bread_s8(impl->rstream)))
{
// -->
if (n == 2 && ch == '>') break;
else
{
// append it
tb_string_chrcat(&impl->element, ch);
if (ch == '-') n++;
else n = 0;
}
}
// update event
if (ch != '\0') impl->event = TB_XML_READER_EVENT_COMMENT;
}
else impl->event = TB_XML_READER_EVENT_COMMENT;
}
// is cdata: <![CDATA[ text ]]>
else if (size >= 8 && !tb_strnicmp(element, "![CDATA[", 8))
{
if (element[size - 2] != ']' || element[size - 1] != ']')
{
// patch '>'
tb_string_chrcat(&impl->element, '>');
// seek to cdata end
tb_char_t ch = '\0';
tb_int_t n = 0;
while ((ch = tb_stream_bread_s8(impl->rstream)))
{
// ]]>
if (n == 2 && ch == '>') break;
else
{
// append it
tb_string_chrcat(&impl->element, ch);
if (ch == ']') n++;
else n = 0;
}
}
// update event
if (ch != '\0') impl->event = TB_XML_READER_EVENT_CDATA;
}
else impl->event = TB_XML_READER_EVENT_CDATA;
}
// is empty element: <name/>
else if (size > 1 && element[size - 1] == '/')
{
// update event
impl->event = TB_XML_READER_EVENT_ELEMENT_EMPTY;
}
// is element begin: <name>
else
{
// update event
impl->event = TB_XML_READER_EVENT_ELEMENT_BEG;
// enter
impl->level++;
}
// trace
// tb_trace_d("<%s>", element);
}
// is text: <> text </>
else if (*pc)
{
// parse text: <> ... <>
tb_char_t const* text = tb_xml_reader_text_parse(impl);
if (text && tb_string_cstrcmp(&impl->text, "\r\n") && tb_string_cstrcmp(&impl->text, "\n"))
impl->event = TB_XML_READER_EVENT_TEXT;
// trace
// tb_trace_d("%s", text);
}
else
{
// skip the invalid character
if (!tb_stream_skip(impl->rstream, 1)) break;
}
}
// ok?
return impl->event;
}
/* //////////////////////////////////////////////////////////////////////////////////////
* main
*/
tb_int_t tb_demo_stream_main(tb_int_t argc, tb_char_t** argv)
{
// done
tb_option_ref_t option = tb_null;
tb_stream_ref_t istream = tb_null;
tb_stream_ref_t ostream = tb_null;
tb_stream_ref_t pstream = tb_null;
do
{
// init option
option = tb_option_init("stream", "the stream demo", g_options);
tb_assert_and_check_break(option);
// done option
if (tb_option_done(option, argc - 1, &argv[1]))
{
// debug & verbose
tb_bool_t debug = tb_option_find(option, "debug");
tb_bool_t verbose = tb_option_find(option, "no-verbose")? tb_false : tb_true;
// done url
if (tb_option_find(option, "url"))
{
// init istream
istream = tb_stream_init_from_url(tb_option_item_cstr(option, "url"));
tb_assert_and_check_break(istream);
// ctrl http
if (tb_stream_type(istream) == TB_STREAM_TYPE_HTTP)
{
// enable gzip?
if (tb_option_find(option, "gzip"))
{
// auto unzip
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_AUTO_UNZIP, 1)) break;
// need gzip
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_HEAD, "Accept-Encoding", "gzip,deflate")) break;
}
// enable debug?
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_HEAD_FUNC, debug? tb_demo_stream_head_func : tb_null)) break;
// custem header?
if (tb_option_find(option, "header"))
{
// init
tb_string_t key;
tb_string_t val;
tb_string_init(&key);
tb_string_init(&val);
// done
tb_bool_t k = tb_true;
tb_char_t const* p = tb_option_item_cstr(option, "header");
while (*p)
{
// is key?
if (k)
{
if (*p != ':' && !tb_isspace(*p)) tb_string_chrcat(&key, *p++);
else if (*p == ':')
{
// skip ':'
p++;
// skip space
while (*p && tb_isspace(*p)) p++;
// is val now
k = tb_false;
}
else p++;
}
// is val?
else
{
if (*p != ';') tb_string_chrcat(&val, *p++);
else
{
// skip ';'
p++;
// skip space
while (*p && tb_isspace(*p)) p++;
// set header
if (tb_string_size(&key) && tb_string_size(&val))
{
if (debug) tb_printf("header: %s: %s\n", tb_string_cstr(&key), tb_string_cstr(&val));
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_HEAD, tb_string_cstr(&key), tb_string_cstr(&val))) break;
}
// is key now
k = tb_true;
// clear key & val
tb_string_clear(&key);
tb_string_clear(&val);
}
}
}
// set header
if (tb_string_size(&key) && tb_string_size(&val))
{
if (debug) tb_printf("header: %s: %s\n", tb_string_cstr(&key), tb_string_cstr(&val));
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_HEAD, tb_string_cstr(&key), tb_string_cstr(&val))) break;
}
// exit
tb_string_exit(&key);
tb_string_exit(&val);
}
// keep alive?
if (tb_option_find(option, "keep-alive"))
{
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_HEAD, "Connection", "keep-alive")) break;
}
// post-data?
if (tb_option_find(option, "post-data"))
{
tb_char_t const* post_data = tb_option_item_cstr(option, "post-data");
tb_hize_t post_size = tb_strlen(post_data);
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_METHOD, TB_HTTP_METHOD_POST)) break;
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_POST_DATA, post_data, post_size)) break;
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_POST_FUNC, tb_demo_http_post_func)) break;
if (debug) tb_printf("post: %llu\n", post_size);
}
// post-file?
else if (tb_option_find(option, "post-file"))
{
tb_char_t const* url = tb_option_item_cstr(option, "post-file");
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_METHOD, TB_HTTP_METHOD_POST)) break;
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_POST_URL, url)) break;
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_POST_FUNC, tb_demo_http_post_func)) break;
if (debug) tb_printf("post: %s\n", url);
}
}
// set range
if (tb_option_find(option, "range"))
{
tb_char_t const* p = tb_option_item_cstr(option, "range");
if (p)
{
// the bof
tb_hize_t eof = 0;
tb_hize_t bof = tb_atoll(p);
while (*p && tb_isdigit(*p)) p++;
if (*p == '-')
{
p++;
eof = tb_atoll(p);
}
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_HTTP_SET_RANGE, bof, eof)) break;
}
}
// set timeout
if (tb_option_find(option, "timeout"))
{
tb_size_t timeout = tb_option_item_uint32(option, "timeout");
if (!tb_stream_ctrl(istream, TB_STREAM_CTRL_SET_TIMEOUT, timeout)) break;
}
// print verbose info
if (verbose) tb_printf("open: %s: ..\n", tb_option_item_cstr(option, "url"));
// open istream
if (!tb_stream_open(istream))
{
// print verbose info
if (verbose) tb_printf("open: %s\n", tb_state_cstr(tb_stream_state(istream)));
break;
}
// print verbose info
if (verbose) tb_printf("open: ok\n");
// init ostream
if (tb_option_find(option, "more0"))
{
// the path
tb_char_t const* path = tb_option_item_cstr(option, "more0");
// init
ostream = tb_stream_init_from_file(path, TB_FILE_MODE_RW | TB_FILE_MODE_CREAT | TB_FILE_MODE_BINARY | TB_FILE_MODE_TRUNC);
// print verbose info
if (verbose) tb_printf("save: %s\n", path);
}
else
{
// the name
tb_char_t const* name = tb_strrchr(tb_option_item_cstr(option, "url"), '/');
if (!name) name = tb_strrchr(tb_option_item_cstr(option, "url"), '\\');
if (!name) name = "/stream.file";
// the path
tb_char_t path[TB_PATH_MAXN] = {0};
if (tb_directory_current(path, TB_PATH_MAXN))
tb_strcat(path, name);
else break;
// init file
ostream = tb_stream_init_from_file(path, TB_FILE_MODE_RW | TB_FILE_MODE_CREAT | TB_FILE_MODE_BINARY | TB_FILE_MODE_TRUNC);
// print verbose info
if (verbose) tb_printf("save: %s\n", path);
}
tb_assert_and_check_break(ostream);
// the limit rate
tb_size_t limitrate = 0;
if (tb_option_find(option, "limitrate"))
limitrate = tb_option_item_uint32(option, "limitrate");
// save it
tb_hong_t save = 0;
tb_demo_context_t context = {0};
context.verbose = verbose;
if ((save = tb_transfer_done(istream, ostream, limitrate, tb_demo_stream_save_func, &context)) < 0) break;
}
else tb_option_help(option);
}
else tb_option_help(option);
} while (0);
// exit pstream
if (pstream) tb_stream_exit(pstream);
pstream = tb_null;
// exit istream
if (istream) tb_stream_exit(istream);
istream = tb_null;
// exit ostream
if (ostream) tb_stream_exit(ostream);
ostream = tb_null;
// exit option
if (option) tb_option_exit(option);
option = tb_null;
return 0;
}
/* chunked_data
*
* head data tail
* ea5\r\n ..........\r\n e65\r\n..............\r\n 0\r\n\r\n
* ---------------------- ------------------------- ---------
* chunk0 chunk1 end
*/
static tb_long_t tb_filter_chunked_spak(tb_filter_t* filter, tb_static_stream_ref_t istream, tb_static_stream_ref_t ostream, tb_long_t sync)
{
// check
tb_filter_chunked_t* cfilter = tb_filter_chunked_cast(filter);
tb_assert_and_check_return_val(cfilter && istream && ostream, -1);
tb_assert_and_check_return_val(tb_static_stream_valid(istream) && tb_static_stream_valid(ostream), -1);
// the idata
tb_byte_t const* ip = tb_static_stream_pos(istream);
tb_byte_t const* ie = tb_static_stream_end(istream);
// trace
tb_trace_d("[%p]: isize: %lu, beof: %d", cfilter, tb_static_stream_size(istream), filter->beof);
// find the eof: '\r\n 0\r\n\r\n'
if ( !filter->beof
&& ip + 6 < ie
&& ie[-7] == '\r'
&& ie[-6] == '\n'
&& ie[-5] == '0'
&& ie[-4] == '\r'
&& ie[-3] == '\n'
&& ie[-2] == '\r'
&& ie[-1] == '\n')
{
// is eof
filter->beof = tb_true;
}
// the odata
tb_byte_t* op = (tb_byte_t*)tb_static_stream_pos(ostream);
tb_byte_t* oe = (tb_byte_t*)tb_static_stream_end(ostream);
tb_byte_t* ob = op;
// parse chunked head and chunked tail
if (!cfilter->size || cfilter->read >= cfilter->size)
{
// walk
while (ip < ie)
{
// the charactor
tb_char_t ch = *ip++;
// trace
tb_trace_d("[%p]: character: %x", cfilter, ch);
// check
tb_assert_and_check_return_val(ch, -1);
// append char to line
if (ch != '\n') tb_string_chrcat(&cfilter->line, ch);
// is line end?
else
{
// check
tb_char_t const* pb = tb_string_cstr(&cfilter->line);
tb_size_t pn = tb_string_size(&cfilter->line);
tb_assert_and_check_return_val(pb, -1);
// trace
tb_trace_d("[%p]: line: %s", cfilter, tb_string_cstr(&cfilter->line));
// strip '\r' if exists
if (pb[pn - 1] == '\r') tb_string_strip(&cfilter->line, pn - 1);
// is chunked tail? only "\r\n"
if (!tb_string_size(&cfilter->line))
{
// reset size
cfilter->read = 0;
cfilter->size = 0;
// trace
tb_trace_d("[%p]: tail", cfilter);
// continue
continue ;
}
// is chunked head? parse size
else
{
// parse size
cfilter->size = tb_s16tou32(pb);
// trace
tb_trace_d("[%p]: size: %lu", cfilter, cfilter->size);
// clear data
tb_string_clear(&cfilter->line);
// is eof? "0\r\n\r\n"
if (!cfilter->size)
{
// trace
tb_trace_d("[%p]: eof", cfilter);
// is eof
filter->beof = tb_true;
// continue to spak the end data
continue ;
}
// ok
break;
}
}
}
}
// check
tb_assert_and_check_return_val(cfilter->read <= cfilter->size, -1);
// read chunked data
tb_size_t size = tb_min3(ie - ip, oe - op, cfilter->size - cfilter->read);
if (size)
{
// copy data
tb_memcpy((tb_byte_t*)op, ip, size);
ip += size;
op += size;
// update read
cfilter->read += size;
}
// update stream
tb_static_stream_goto(istream, (tb_byte_t*)ip);
tb_static_stream_goto(ostream, (tb_byte_t*)op);
// trace
tb_trace_d("[%p]: read: %lu, size: %lu, beof: %u, ileft: %lu", cfilter, cfilter->read, cfilter->size, filter->beof, tb_static_stream_left(istream));
// ok
return (op - ob);
}
