2018-12-31 20:04:05 +01:00
|
|
|
//
|
2024-01-01 01:07:21 +01:00
|
|
|
// Copyright Aliaksei Levin (levlam@telegram.org), Arseny Smirnov (arseny30@gmail.com) 2014-2024
|
2018-12-31 20:04:05 +01:00
|
|
|
//
|
|
|
|
// Distributed under the Boost Software License, Version 1.0. (See accompanying
|
|
|
|
// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
|
|
|
|
//
|
|
|
|
#include "td/net/HttpReader.h"
|
|
|
|
|
|
|
|
#include "td/utils/filesystem.h"
|
|
|
|
#include "td/utils/find_boundary.h"
|
|
|
|
#include "td/utils/format.h"
|
|
|
|
#include "td/utils/Gzip.h"
|
|
|
|
#include "td/utils/HttpUrl.h"
|
|
|
|
#include "td/utils/JsonBuilder.h"
|
|
|
|
#include "td/utils/logging.h"
|
|
|
|
#include "td/utils/misc.h"
|
|
|
|
#include "td/utils/Parser.h"
|
|
|
|
#include "td/utils/PathView.h"
|
|
|
|
#include "td/utils/port/path.h"
|
2021-05-17 14:21:11 +02:00
|
|
|
#include "td/utils/SliceBuilder.h"
|
2018-12-31 20:04:05 +01:00
|
|
|
|
2021-10-20 00:27:02 +02:00
|
|
|
#include <cstddef>
|
2018-12-31 20:04:05 +01:00
|
|
|
#include <cstring>
|
|
|
|
|
|
|
|
namespace td {
|
|
|
|
|
|
|
|
constexpr const char HttpReader::TEMP_DIRECTORY_PREFIX[];
|
|
|
|
|
|
|
|
void HttpReader::init(ChainBufferReader *input, size_t max_post_size, size_t max_files) {
|
|
|
|
input_ = input;
|
2020-01-19 01:02:56 +01:00
|
|
|
state_ = State::ReadHeaders;
|
2018-12-31 20:04:05 +01:00
|
|
|
headers_read_length_ = 0;
|
2023-11-25 00:45:04 +01:00
|
|
|
content_length_ = -1;
|
2018-12-31 20:04:05 +01:00
|
|
|
query_ = nullptr;
|
|
|
|
max_post_size_ = max_post_size;
|
|
|
|
max_files_ = max_files;
|
|
|
|
total_parameters_length_ = 0;
|
|
|
|
total_headers_length_ = 0;
|
|
|
|
}
|
|
|
|
|
2020-07-23 18:47:12 +02:00
|
|
|
Result<size_t> HttpReader::read_next(HttpQuery *query, bool can_be_slow) {
|
2018-12-31 20:04:05 +01:00
|
|
|
if (query_ != query) {
|
|
|
|
CHECK(query_ == nullptr);
|
|
|
|
query_ = query;
|
|
|
|
}
|
2023-11-25 00:34:37 +01:00
|
|
|
|
|
|
|
auto r_size = do_read_next(can_be_slow);
|
|
|
|
if (state_ != State::ReadHeaders && flow_sink_.is_ready() && r_size.is_ok() && r_size.ok() > 0) {
|
|
|
|
CHECK(flow_sink_.status().is_ok());
|
|
|
|
return Status::Error(400, "Bad Request: unexpected end of request content");
|
|
|
|
}
|
|
|
|
return r_size;
|
|
|
|
}
|
|
|
|
|
|
|
|
Result<size_t> HttpReader::do_read_next(bool can_be_slow) {
|
2018-12-31 20:04:05 +01:00
|
|
|
size_t need_size = input_->size() + 1;
|
|
|
|
while (true) {
|
2020-01-19 01:02:56 +01:00
|
|
|
if (state_ != State::ReadHeaders) {
|
2020-07-22 20:52:00 +02:00
|
|
|
gzip_flow_.wakeup();
|
2018-12-31 20:04:05 +01:00
|
|
|
flow_source_.wakeup();
|
|
|
|
if (flow_sink_.is_ready() && flow_sink_.status().is_error()) {
|
2018-08-29 00:57:09 +02:00
|
|
|
if (!temp_file_.empty()) {
|
|
|
|
clean_temporary_file();
|
|
|
|
}
|
2018-07-08 01:47:46 +02:00
|
|
|
return Status::Error(400, PSLICE() << "Bad Request: " << flow_sink_.status().message());
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
need_size = flow_source_.get_need_size();
|
|
|
|
if (need_size == 0) {
|
|
|
|
need_size = input_->size() + 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
switch (state_) {
|
2020-01-19 01:02:56 +01:00
|
|
|
case State::ReadHeaders: {
|
2018-12-31 20:04:05 +01:00
|
|
|
auto result = split_header();
|
|
|
|
if (result.is_error() || result.ok() != 0) {
|
|
|
|
return result;
|
|
|
|
}
|
2023-11-25 00:45:04 +01:00
|
|
|
if (transfer_encoding_.empty() && content_length_ <= 0) {
|
2018-12-31 20:04:05 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
flow_source_ = ByteFlowSource(input_);
|
|
|
|
ByteFlowInterface *source = &flow_source_;
|
|
|
|
if (transfer_encoding_.empty()) {
|
2023-11-25 00:45:04 +01:00
|
|
|
content_length_flow_ = HttpContentLengthByteFlow(narrow_cast<size_t>(content_length_));
|
2018-12-31 20:04:05 +01:00
|
|
|
*source >> content_length_flow_;
|
|
|
|
source = &content_length_flow_;
|
|
|
|
} else if (transfer_encoding_ == "chunked") {
|
|
|
|
chunked_flow_ = HttpChunkedByteFlow();
|
|
|
|
*source >> chunked_flow_;
|
|
|
|
source = &chunked_flow_;
|
|
|
|
} else {
|
|
|
|
LOG(ERROR) << "Unsupported " << tag("transfer-encoding", transfer_encoding_);
|
|
|
|
return Status::Error(501, "Unimplemented: unsupported transfer-encoding");
|
|
|
|
}
|
|
|
|
|
2022-11-29 11:15:33 +01:00
|
|
|
if (content_encoding_.empty() || content_encoding_ == "none") {
|
2018-12-31 20:04:05 +01:00
|
|
|
} else if (content_encoding_ == "gzip" || content_encoding_ == "deflate") {
|
2020-01-19 01:02:56 +01:00
|
|
|
gzip_flow_ = GzipByteFlow(Gzip::Mode::Decode);
|
2020-07-22 20:52:00 +02:00
|
|
|
GzipByteFlow::Options options;
|
|
|
|
options.write_watermark.low = 0;
|
2021-03-20 22:46:36 +01:00
|
|
|
options.write_watermark.high = max(max_post_size_, MAX_TOTAL_PARAMETERS_LENGTH + 1);
|
2020-07-22 20:52:00 +02:00
|
|
|
gzip_flow_.set_options(options);
|
2020-07-23 15:48:17 +02:00
|
|
|
gzip_flow_.set_max_output_size(MAX_CONTENT_SIZE);
|
2018-12-31 20:04:05 +01:00
|
|
|
*source >> gzip_flow_;
|
|
|
|
source = &gzip_flow_;
|
|
|
|
} else {
|
2018-08-28 04:52:43 +02:00
|
|
|
LOG(WARNING) << "Unsupported " << tag("content-encoding", content_encoding_);
|
2018-12-31 20:04:05 +01:00
|
|
|
return Status::Error(415, "Unsupported Media Type: unsupported content-encoding");
|
|
|
|
}
|
|
|
|
|
|
|
|
flow_sink_ = ByteFlowSink();
|
|
|
|
*source >> flow_sink_;
|
|
|
|
content_ = flow_sink_.get_output();
|
|
|
|
|
2023-11-25 00:45:04 +01:00
|
|
|
if (content_length_ >= static_cast<int64>(MAX_CONTENT_SIZE)) {
|
2018-12-31 20:04:05 +01:00
|
|
|
return Status::Error(413, PSLICE() << "Request Entity Too Large: content length is " << content_length_);
|
|
|
|
}
|
|
|
|
|
2018-11-01 18:08:20 +01:00
|
|
|
if (content_type_lowercased_.find("multipart/form-data") != string::npos) {
|
2020-01-19 01:02:56 +01:00
|
|
|
state_ = State::ReadMultipartFormData;
|
2018-12-31 20:04:05 +01:00
|
|
|
|
|
|
|
const char *p = std::strstr(content_type_lowercased_.c_str(), "boundary");
|
|
|
|
if (p == nullptr) {
|
|
|
|
return Status::Error(400, "Bad Request: boundary not found");
|
|
|
|
}
|
|
|
|
p += 8;
|
2021-10-20 00:27:02 +02:00
|
|
|
std::ptrdiff_t offset = p - content_type_lowercased_.c_str();
|
2018-12-31 20:04:05 +01:00
|
|
|
p = static_cast<const char *>(
|
|
|
|
std::memchr(content_type_.begin() + offset, '=', content_type_.size() - offset));
|
|
|
|
if (p == nullptr) {
|
|
|
|
return Status::Error(400, "Bad Request: boundary value not found");
|
|
|
|
}
|
|
|
|
p++;
|
2021-10-20 00:27:02 +02:00
|
|
|
auto end_p = static_cast<const char *>(std::memchr(p, ';', content_type_.end() - p));
|
2018-12-31 20:04:05 +01:00
|
|
|
if (end_p == nullptr) {
|
|
|
|
end_p = content_type_.end();
|
|
|
|
}
|
|
|
|
if (*p == '"' && p + 1 < end_p && end_p[-1] == '"') {
|
|
|
|
p++;
|
|
|
|
end_p--;
|
|
|
|
}
|
|
|
|
|
2020-05-19 14:11:21 +02:00
|
|
|
CHECK(p != nullptr);
|
2018-12-31 20:04:05 +01:00
|
|
|
Slice boundary(p, static_cast<size_t>(end_p - p));
|
|
|
|
if (boundary.empty() || boundary.size() > MAX_BOUNDARY_LENGTH) {
|
|
|
|
return Status::Error(400, "Bad Request: boundary too big or empty");
|
|
|
|
}
|
|
|
|
|
|
|
|
boundary_ = "\r\n--" + boundary.str();
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::SkipPrologue;
|
2018-12-31 20:04:05 +01:00
|
|
|
form_data_read_length_ = 0;
|
|
|
|
form_data_skipped_length_ = 0;
|
2018-11-01 18:08:20 +01:00
|
|
|
} else if (content_type_lowercased_.find("application/x-www-form-urlencoded") != string::npos ||
|
|
|
|
content_type_lowercased_.find("application/json") != string::npos) {
|
2020-01-19 01:02:56 +01:00
|
|
|
state_ = State::ReadArgs;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
|
|
|
form_data_skipped_length_ = 0;
|
2020-01-19 01:02:56 +01:00
|
|
|
state_ = State::ReadContent;
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
continue;
|
|
|
|
}
|
2020-01-19 01:02:56 +01:00
|
|
|
case State::ReadContent: {
|
2018-12-31 20:04:05 +01:00
|
|
|
if (content_->size() > max_post_size_) {
|
2020-01-19 01:02:56 +01:00
|
|
|
state_ = State::ReadContentToFile;
|
2020-07-22 20:52:00 +02:00
|
|
|
GzipByteFlow::Options options;
|
|
|
|
options.write_watermark.low = 4 << 20;
|
2020-07-23 01:12:00 +02:00
|
|
|
options.write_watermark.high = 8 << 20;
|
2020-07-22 20:52:00 +02:00
|
|
|
gzip_flow_.set_options(options);
|
2018-12-31 20:04:05 +01:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (flow_sink_.is_ready()) {
|
|
|
|
CHECK(query_->container_.size() == 1u);
|
|
|
|
query_->container_.emplace_back(content_->cut_head(content_->size()).move_as_buffer_slice());
|
2023-02-02 15:40:47 +01:00
|
|
|
query_->content_ = query_->container_.back().as_mutable_slice();
|
2018-12-31 20:04:05 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return need_size;
|
|
|
|
}
|
2020-01-19 01:02:56 +01:00
|
|
|
case State::ReadContentToFile: {
|
2020-07-23 18:47:12 +02:00
|
|
|
if (!can_be_slow) {
|
|
|
|
return Status::Error("SLOW");
|
|
|
|
}
|
2018-12-31 20:04:05 +01:00
|
|
|
// save content to a file
|
|
|
|
if (temp_file_.empty()) {
|
2022-12-07 11:43:27 +01:00
|
|
|
auto open_status = open_temp_file("file");
|
|
|
|
if (open_status.is_error()) {
|
2018-12-31 20:04:05 +01:00
|
|
|
return Status::Error(500, "Internal Server Error: can't create temporary file");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
auto size = content_->size();
|
2020-07-22 20:52:00 +02:00
|
|
|
bool restart = false;
|
|
|
|
if (size > (1 << 20) || flow_sink_.is_ready()) {
|
2018-12-31 20:04:05 +01:00
|
|
|
TRY_STATUS(save_file_part(content_->cut_head(size).move_as_buffer_slice()));
|
2020-07-22 20:52:00 +02:00
|
|
|
restart = true;
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
if (flow_sink_.is_ready()) {
|
|
|
|
query_->files_.emplace_back("file", "", content_type_.str(), file_size_, temp_file_name_);
|
|
|
|
close_temp_file();
|
|
|
|
break;
|
|
|
|
}
|
2020-07-22 20:52:00 +02:00
|
|
|
if (restart) {
|
|
|
|
continue;
|
|
|
|
}
|
2018-12-31 20:04:05 +01:00
|
|
|
|
|
|
|
return need_size;
|
|
|
|
}
|
2020-01-19 01:02:56 +01:00
|
|
|
case State::ReadArgs: {
|
2018-12-31 20:04:05 +01:00
|
|
|
auto size = content_->size();
|
|
|
|
if (size > MAX_TOTAL_PARAMETERS_LENGTH - total_parameters_length_) {
|
2022-01-24 13:01:35 +01:00
|
|
|
return Status::Error(413, "Request Entity Too Large: too many parameters");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
if (flow_sink_.is_ready()) {
|
|
|
|
query_->container_.emplace_back(content_->cut_head(size).move_as_buffer_slice());
|
|
|
|
Status result;
|
2018-11-01 18:08:20 +01:00
|
|
|
if (content_type_lowercased_.find("application/x-www-form-urlencoded") != string::npos) {
|
2023-02-02 15:40:47 +01:00
|
|
|
result = parse_parameters(query_->container_.back().as_mutable_slice());
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
2023-02-02 15:40:47 +01:00
|
|
|
result = parse_json_parameters(query_->container_.back().as_mutable_slice());
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
if (result.is_error()) {
|
|
|
|
if (result.code() == 413) {
|
|
|
|
return std::move(result);
|
|
|
|
}
|
|
|
|
LOG(INFO) << result.message();
|
|
|
|
}
|
|
|
|
query_->content_ = MutableSlice();
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return need_size;
|
|
|
|
}
|
2020-01-19 01:02:56 +01:00
|
|
|
case State::ReadMultipartFormData: {
|
2021-09-06 19:13:32 +02:00
|
|
|
if (!content_->empty() || flow_sink_.is_ready()) {
|
2020-07-24 14:32:33 +02:00
|
|
|
TRY_RESULT(result, parse_multipart_form_data(can_be_slow));
|
2020-07-19 13:49:12 +02:00
|
|
|
if (result) {
|
|
|
|
break;
|
|
|
|
}
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
return need_size;
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
UNREACHABLE();
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
init(input_, max_post_size_, max_files_);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
// returns Status on wrong request
|
|
|
|
// returns true if parsing has finished
|
|
|
|
// returns false if need more data
|
2020-07-24 14:32:33 +02:00
|
|
|
Result<bool> HttpReader::parse_multipart_form_data(bool can_be_slow) {
|
2018-12-31 20:04:05 +01:00
|
|
|
while (true) {
|
2020-07-19 13:49:12 +02:00
|
|
|
LOG(DEBUG) << "Parsing multipart form data in state " << static_cast<int32>(form_data_parse_state_)
|
|
|
|
<< " with already read length " << form_data_read_length_;
|
2018-12-31 20:04:05 +01:00
|
|
|
switch (form_data_parse_state_) {
|
2020-01-19 01:02:56 +01:00
|
|
|
case FormDataParseState::SkipPrologue:
|
2018-12-31 20:04:05 +01:00
|
|
|
if (find_boundary(content_->clone(), {boundary_.c_str() + 2, boundary_.size() - 2}, form_data_read_length_)) {
|
|
|
|
size_t to_skip = form_data_read_length_ + (boundary_.size() - 2);
|
|
|
|
content_->advance(to_skip);
|
|
|
|
form_data_skipped_length_ += to_skip;
|
|
|
|
form_data_read_length_ = 0;
|
|
|
|
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::ReadPartHeaders;
|
2018-12-31 20:04:05 +01:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
content_->advance(form_data_read_length_);
|
|
|
|
form_data_skipped_length_ += form_data_read_length_;
|
|
|
|
form_data_read_length_ = 0;
|
|
|
|
return false;
|
2020-01-19 01:02:56 +01:00
|
|
|
case FormDataParseState::ReadPartHeaders:
|
2018-12-31 20:04:05 +01:00
|
|
|
if (find_boundary(content_->clone(), "\r\n\r\n", form_data_read_length_)) {
|
|
|
|
total_headers_length_ += form_data_read_length_;
|
|
|
|
if (total_headers_length_ > MAX_TOTAL_HEADERS_LENGTH) {
|
|
|
|
return Status::Error(431, "Request Header Fields Too Large: total headers size exceeded");
|
|
|
|
}
|
|
|
|
if (form_data_read_length_ == 0) {
|
2021-03-21 00:58:29 +01:00
|
|
|
// there are no headers at all
|
2018-12-31 20:04:05 +01:00
|
|
|
return Status::Error(400, "Bad Request: headers in multipart/form-data are empty");
|
|
|
|
}
|
|
|
|
|
|
|
|
content_->advance(2); // "\r\n" after boundary
|
|
|
|
auto headers = content_->cut_head(form_data_read_length_).move_as_buffer_slice();
|
|
|
|
CHECK(headers.as_slice().size() == form_data_read_length_);
|
|
|
|
LOG(DEBUG) << "Parse headers in multipart form data: \"" << headers.as_slice() << "\"";
|
|
|
|
content_->advance(2);
|
|
|
|
|
|
|
|
form_data_skipped_length_ += form_data_read_length_ + 4;
|
|
|
|
form_data_read_length_ = 0;
|
|
|
|
|
|
|
|
field_name_ = MutableSlice();
|
|
|
|
file_field_name_.clear();
|
|
|
|
field_content_type_ = "application/octet-stream";
|
|
|
|
file_name_.clear();
|
2018-10-14 18:08:15 +02:00
|
|
|
has_file_name_ = false;
|
2018-12-31 20:04:05 +01:00
|
|
|
CHECK(temp_file_.empty());
|
|
|
|
temp_file_name_.clear();
|
|
|
|
|
2023-02-02 15:40:47 +01:00
|
|
|
Parser headers_parser(headers.as_mutable_slice());
|
2018-12-31 20:04:05 +01:00
|
|
|
while (headers_parser.status().is_ok() && !headers_parser.data().empty()) {
|
|
|
|
MutableSlice header_name = headers_parser.read_till(':');
|
|
|
|
headers_parser.skip(':');
|
|
|
|
char *header_value_begin = headers_parser.ptr();
|
|
|
|
char *header_value_end;
|
|
|
|
do {
|
|
|
|
headers_parser.read_till('\r');
|
|
|
|
header_value_end = headers_parser.ptr();
|
|
|
|
headers_parser.skip('\r');
|
|
|
|
headers_parser.skip('\n');
|
|
|
|
} while (headers_parser.status().is_ok() &&
|
|
|
|
(headers_parser.peek_char() == ' ' || headers_parser.peek_char() == '\t'));
|
|
|
|
|
|
|
|
MutableSlice header_value(header_value_begin, header_value_end);
|
|
|
|
|
|
|
|
header_name = trim(header_name);
|
|
|
|
header_value = trim(header_value);
|
|
|
|
to_lower_inplace(header_name);
|
|
|
|
|
|
|
|
if (header_name == "content-disposition") {
|
|
|
|
if (header_value.substr(0, 10) != "form-data;") {
|
|
|
|
return Status::Error(400, "Bad Request: expected form-data content disposition");
|
|
|
|
}
|
|
|
|
header_value.remove_prefix(10);
|
|
|
|
while (true) {
|
|
|
|
header_value = trim(header_value);
|
2022-08-18 22:31:14 +02:00
|
|
|
const auto *key_end =
|
2018-12-31 20:04:05 +01:00
|
|
|
static_cast<const char *>(std::memchr(header_value.data(), '=', header_value.size()));
|
|
|
|
if (key_end == nullptr) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
size_t key_size = key_end - header_value.data();
|
2021-03-21 00:58:29 +01:00
|
|
|
auto key = trim(header_value.substr(0, key_size));
|
2018-12-31 20:04:05 +01:00
|
|
|
|
|
|
|
header_value.remove_prefix(key_size + 1);
|
2021-03-21 00:58:29 +01:00
|
|
|
|
|
|
|
while (!header_value.empty() && is_space(header_value[0])) {
|
|
|
|
header_value.remove_prefix(1);
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
2021-03-21 00:58:29 +01:00
|
|
|
|
|
|
|
MutableSlice value;
|
|
|
|
if (!header_value.empty() && header_value[0] == '"') { // quoted-string
|
|
|
|
char *value_end = header_value.data() + 1;
|
|
|
|
const char *pos = value_end;
|
|
|
|
while (true) {
|
|
|
|
if (pos == header_value.data() + header_value.size()) {
|
|
|
|
return Status::Error(400, "Bad Request: unclosed quoted string in Content-Disposition header");
|
|
|
|
}
|
|
|
|
char c = *pos++;
|
|
|
|
if (c == '"') {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (c == '\\') {
|
|
|
|
if (pos == header_value.data() + header_value.size()) {
|
|
|
|
return Status::Error(400, "Bad Request: wrong escape sequence in Content-Disposition header");
|
|
|
|
}
|
|
|
|
c = *pos++;
|
|
|
|
}
|
|
|
|
*value_end++ = c;
|
|
|
|
}
|
|
|
|
value = header_value.substr(1, value_end - header_value.data() - 1);
|
|
|
|
header_value.remove_prefix(pos - header_value.data());
|
|
|
|
|
|
|
|
while (!header_value.empty() && is_space(header_value[0])) {
|
|
|
|
header_value.remove_prefix(1);
|
|
|
|
}
|
|
|
|
if (!header_value.empty()) {
|
|
|
|
if (header_value[0] != ';') {
|
|
|
|
return Status::Error(400, "Bad Request: expected ';' in Content-Disposition header");
|
|
|
|
}
|
|
|
|
header_value.remove_prefix(1);
|
|
|
|
}
|
|
|
|
} else { // token
|
|
|
|
auto value_end =
|
|
|
|
static_cast<const char *>(std::memchr(header_value.data(), ';', header_value.size()));
|
|
|
|
if (value_end != nullptr) {
|
|
|
|
auto value_size = static_cast<size_t>(value_end - header_value.data());
|
|
|
|
value = trim(header_value.substr(0, value_size));
|
|
|
|
header_value.remove_prefix(value_size + 1);
|
|
|
|
} else {
|
|
|
|
value = trim(header_value);
|
|
|
|
header_value = MutableSlice();
|
|
|
|
}
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
2023-02-06 20:44:01 +01:00
|
|
|
value = url_decode_inplace(value, false);
|
2018-12-31 20:04:05 +01:00
|
|
|
|
|
|
|
if (key == "name") {
|
|
|
|
field_name_ = value;
|
|
|
|
} else if (key == "filename") {
|
|
|
|
file_name_ = value.str();
|
2018-10-14 18:08:15 +02:00
|
|
|
has_file_name_ = true;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
|
|
|
// ignore unknown parts of header
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else if (header_name == "content-type") {
|
|
|
|
field_content_type_ = header_value.str();
|
|
|
|
} else {
|
|
|
|
// ignore unknown header
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (headers_parser.status().is_error()) {
|
|
|
|
return Status::Error(400, "Bad Request: can't parse form data headers");
|
|
|
|
}
|
|
|
|
|
|
|
|
if (field_name_.empty()) {
|
|
|
|
return Status::Error(400, "Bad Request: field name in multipart/form-data not found");
|
|
|
|
}
|
|
|
|
|
2018-10-14 18:08:15 +02:00
|
|
|
if (has_file_name_) {
|
2018-12-31 20:04:05 +01:00
|
|
|
// file
|
|
|
|
if (query_->files_.size() == max_files_) {
|
2022-01-24 13:01:35 +01:00
|
|
|
return Status::Error(413, "Request Entity Too Large: too many files attached");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
// don't need to save headers for files
|
|
|
|
file_field_name_ = field_name_.str();
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::ReadFile;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
|
|
|
// save headers for query parameters. They contain header names
|
|
|
|
query_->container_.push_back(std::move(headers));
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::ReadPartValue;
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (total_headers_length_ + form_data_read_length_ > MAX_TOTAL_HEADERS_LENGTH) {
|
|
|
|
return Status::Error(431, "Request Header Fields Too Large: total headers size exceeded");
|
|
|
|
}
|
|
|
|
return false;
|
2020-01-19 01:02:56 +01:00
|
|
|
case FormDataParseState::ReadPartValue:
|
2018-12-31 20:04:05 +01:00
|
|
|
if (find_boundary(content_->clone(), boundary_, form_data_read_length_)) {
|
|
|
|
if (total_parameters_length_ + form_data_read_length_ > MAX_TOTAL_PARAMETERS_LENGTH) {
|
2022-01-24 13:01:35 +01:00
|
|
|
return Status::Error(413, "Request Entity Too Large: too many parameters in form data");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
query_->container_.emplace_back(content_->cut_head(form_data_read_length_).move_as_buffer_slice());
|
2023-02-02 15:40:47 +01:00
|
|
|
MutableSlice value = query_->container_.back().as_mutable_slice();
|
2018-12-31 20:04:05 +01:00
|
|
|
content_->advance(boundary_.size());
|
|
|
|
form_data_skipped_length_ += form_data_read_length_ + boundary_.size();
|
|
|
|
form_data_read_length_ = 0;
|
|
|
|
|
|
|
|
if (begins_with(field_content_type_, "application/x-www-form-urlencoded")) {
|
|
|
|
// treat value as ordinary parameters
|
|
|
|
auto result = parse_parameters(value);
|
|
|
|
if (result.is_error()) {
|
|
|
|
return std::move(result);
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
total_parameters_length_ += form_data_read_length_;
|
|
|
|
LOG(DEBUG) << "Get ordinary parameter in multipart form data: \"" << field_name_ << "\": \"" << value
|
|
|
|
<< "\"";
|
|
|
|
query_->args_.emplace_back(field_name_, value);
|
|
|
|
}
|
|
|
|
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::CheckForLastBoundary;
|
2018-12-31 20:04:05 +01:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
CHECK(content_->size() < form_data_read_length_ + boundary_.size());
|
|
|
|
|
|
|
|
if (total_parameters_length_ + form_data_read_length_ > MAX_TOTAL_PARAMETERS_LENGTH) {
|
2022-01-24 13:01:35 +01:00
|
|
|
return Status::Error(413, "Request Entity Too Large: too many parameters in form data");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
return false;
|
2020-01-19 01:02:56 +01:00
|
|
|
case FormDataParseState::ReadFile: {
|
2020-07-24 14:32:33 +02:00
|
|
|
if (!can_be_slow) {
|
|
|
|
return Status::Error("SLOW");
|
|
|
|
}
|
2022-12-07 11:43:27 +01:00
|
|
|
if (temp_file_.empty()) {
|
|
|
|
auto open_status = open_temp_file(file_name_);
|
|
|
|
if (open_status.is_error()) {
|
|
|
|
return Status::Error(500, "Internal Server Error: can't create temporary file");
|
|
|
|
}
|
|
|
|
}
|
2018-12-31 20:04:05 +01:00
|
|
|
if (find_boundary(content_->clone(), boundary_, form_data_read_length_)) {
|
|
|
|
auto file_part = content_->cut_head(form_data_read_length_).move_as_buffer_slice();
|
|
|
|
content_->advance(boundary_.size());
|
|
|
|
form_data_skipped_length_ += form_data_read_length_ + boundary_.size();
|
|
|
|
form_data_read_length_ = 0;
|
|
|
|
|
|
|
|
TRY_STATUS(save_file_part(std::move(file_part)));
|
|
|
|
|
|
|
|
query_->files_.emplace_back(file_field_name_, file_name_, field_content_type_, file_size_, temp_file_name_);
|
|
|
|
close_temp_file();
|
|
|
|
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::CheckForLastBoundary;
|
2018-12-31 20:04:05 +01:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
// TODO optimize?
|
|
|
|
auto file_part = content_->cut_head(form_data_read_length_).move_as_buffer_slice();
|
|
|
|
form_data_skipped_length_ += form_data_read_length_;
|
|
|
|
form_data_read_length_ = 0;
|
|
|
|
CHECK(content_->size() < boundary_.size());
|
|
|
|
|
|
|
|
TRY_STATUS(save_file_part(std::move(file_part)));
|
|
|
|
return false;
|
|
|
|
}
|
2020-01-19 01:02:56 +01:00
|
|
|
case FormDataParseState::CheckForLastBoundary: {
|
2018-12-31 20:04:05 +01:00
|
|
|
if (content_->size() < 2) {
|
|
|
|
// need more data
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
auto range = content_->clone();
|
|
|
|
char x[2];
|
|
|
|
range.advance(2, {x, 2});
|
|
|
|
if (x[0] == '-' && x[1] == '-') {
|
|
|
|
content_->advance(2);
|
|
|
|
form_data_skipped_length_ += 2;
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::SkipEpilogue;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
2020-01-19 01:02:56 +01:00
|
|
|
form_data_parse_state_ = FormDataParseState::ReadPartHeaders;
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
continue;
|
|
|
|
}
|
2020-01-19 01:02:56 +01:00
|
|
|
case FormDataParseState::SkipEpilogue: {
|
2018-12-31 20:04:05 +01:00
|
|
|
size_t size = content_->size();
|
|
|
|
LOG(DEBUG) << "Skipping epilogue. Have " << size << " bytes";
|
|
|
|
content_->advance(size);
|
|
|
|
form_data_skipped_length_ += size;
|
|
|
|
// TODO(now): check if form_data_skipped_length is too big
|
|
|
|
return flow_sink_.is_ready();
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
UNREACHABLE();
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
Result<size_t> HttpReader::split_header() {
|
|
|
|
if (find_boundary(input_->clone(), "\r\n\r\n", headers_read_length_)) {
|
|
|
|
query_->container_.clear();
|
|
|
|
auto a = input_->cut_head(headers_read_length_ + 2);
|
|
|
|
auto b = a.move_as_buffer_slice();
|
|
|
|
query_->container_.emplace_back(std::move(b));
|
|
|
|
// query_->container_.emplace_back(input_->cut_head(headers_read_length_ + 2).move_as_buffer_slice());
|
|
|
|
CHECK(query_->container_.back().size() == headers_read_length_ + 2);
|
|
|
|
input_->advance(2);
|
|
|
|
total_headers_length_ = headers_read_length_;
|
2023-02-02 15:40:47 +01:00
|
|
|
auto status = parse_head(query_->container_.back().as_mutable_slice());
|
2018-12-31 20:04:05 +01:00
|
|
|
if (status.is_error()) {
|
|
|
|
return std::move(status);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (input_->size() > MAX_TOTAL_HEADERS_LENGTH) {
|
|
|
|
return Status::Error(431, "Request Header Fields Too Large: total headers size exceeded");
|
|
|
|
}
|
|
|
|
return input_->size() + 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
void HttpReader::process_header(MutableSlice header_name, MutableSlice header_value) {
|
|
|
|
header_name = trim(header_name);
|
|
|
|
header_value = trim(header_value); // TODO need to remove "\r\n" from value
|
|
|
|
to_lower_inplace(header_name);
|
2019-02-21 16:58:20 +01:00
|
|
|
LOG(DEBUG) << "Process header [" << header_name << "=>" << header_value << "]";
|
2018-12-31 20:04:05 +01:00
|
|
|
query_->headers_.emplace_back(header_name, header_value);
|
|
|
|
if (header_name == "content-length") {
|
2022-08-14 12:30:25 +02:00
|
|
|
auto content_length = to_integer<uint64>(header_value);
|
|
|
|
if (content_length > MAX_CONTENT_SIZE) {
|
|
|
|
content_length = MAX_CONTENT_SIZE;
|
|
|
|
}
|
2023-11-25 00:45:04 +01:00
|
|
|
content_length_ = static_cast<int64>(content_length);
|
2018-12-31 20:04:05 +01:00
|
|
|
} else if (header_name == "connection") {
|
|
|
|
to_lower_inplace(header_value);
|
|
|
|
if (header_value == "close") {
|
|
|
|
query_->keep_alive_ = false;
|
2022-11-22 13:46:53 +01:00
|
|
|
} else {
|
|
|
|
query_->keep_alive_ = true;
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
} else if (header_name == "content-type") {
|
|
|
|
content_type_ = header_value;
|
|
|
|
content_type_lowercased_ = header_value.str();
|
|
|
|
to_lower_inplace(content_type_lowercased_);
|
|
|
|
} else if (header_name == "content-encoding") {
|
|
|
|
to_lower_inplace(header_value);
|
|
|
|
content_encoding_ = header_value;
|
|
|
|
} else if (header_name == "transfer-encoding") {
|
|
|
|
to_lower_inplace(header_value);
|
|
|
|
transfer_encoding_ = header_value;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
Status HttpReader::parse_url(MutableSlice url) {
|
|
|
|
size_t url_path_size = 0;
|
|
|
|
while (url_path_size < url.size() && url[url_path_size] != '?' && url[url_path_size] != '#') {
|
|
|
|
url_path_size++;
|
|
|
|
}
|
|
|
|
|
2020-09-14 17:58:39 +02:00
|
|
|
query_->url_path_ = url_decode_inplace({url.data(), url_path_size}, false);
|
2018-12-31 20:04:05 +01:00
|
|
|
|
|
|
|
if (url_path_size == url.size() || url[url_path_size] != '?') {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
return parse_parameters(url.substr(url_path_size + 1));
|
|
|
|
}
|
|
|
|
|
|
|
|
Status HttpReader::parse_parameters(MutableSlice parameters) {
|
|
|
|
total_parameters_length_ += parameters.size();
|
|
|
|
if (total_parameters_length_ > MAX_TOTAL_PARAMETERS_LENGTH) {
|
2022-01-24 13:01:35 +01:00
|
|
|
return Status::Error(413, "Request Entity Too Large: too many parameters");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
LOG(DEBUG) << "Parse parameters: \"" << parameters << "\"";
|
|
|
|
|
|
|
|
Parser parser(parameters);
|
|
|
|
while (!parser.data().empty()) {
|
|
|
|
auto key_value = parser.read_till_nofail('&');
|
|
|
|
parser.skip_nofail('&');
|
|
|
|
Parser kv_parser(key_value);
|
2020-09-14 17:58:39 +02:00
|
|
|
auto key = url_decode_inplace(kv_parser.read_till_nofail('='), true);
|
2018-12-31 20:04:05 +01:00
|
|
|
kv_parser.skip_nofail('=');
|
2020-09-14 17:58:39 +02:00
|
|
|
auto value = url_decode_inplace(kv_parser.data(), true);
|
2018-12-31 20:04:05 +01:00
|
|
|
query_->args_.emplace_back(key, value);
|
|
|
|
}
|
|
|
|
|
|
|
|
CHECK(parser.status().is_ok());
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
|
|
|
Status HttpReader::parse_json_parameters(MutableSlice parameters) {
|
|
|
|
if (parameters.empty()) {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
|
|
|
total_parameters_length_ += parameters.size();
|
|
|
|
if (total_parameters_length_ > MAX_TOTAL_PARAMETERS_LENGTH) {
|
2022-01-24 13:01:35 +01:00
|
|
|
return Status::Error(413, "Request Entity Too Large: too many parameters");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
2019-08-26 04:06:00 +02:00
|
|
|
LOG(DEBUG) << "Parse JSON parameters: \"" << parameters << "\"";
|
2018-12-31 20:04:05 +01:00
|
|
|
|
|
|
|
Parser parser(parameters);
|
|
|
|
parser.skip_whitespaces();
|
2019-11-19 17:01:12 +01:00
|
|
|
if (parser.peek_char() == '"') {
|
|
|
|
auto r_value = json_string_decode(parser);
|
|
|
|
if (r_value.is_error()) {
|
|
|
|
return Status::Error(400, PSLICE() << "Bad Request: can't parse string content: " << r_value.error().message());
|
|
|
|
}
|
|
|
|
if (!parser.empty()) {
|
|
|
|
return Status::Error(400, "Bad Request: extra data after string");
|
|
|
|
}
|
2022-10-27 11:53:53 +02:00
|
|
|
query_->container_.emplace_back("content");
|
2023-02-02 15:40:47 +01:00
|
|
|
query_->args_.emplace_back(query_->container_.back().as_mutable_slice(), r_value.move_as_ok());
|
2019-11-19 17:01:12 +01:00
|
|
|
return Status::OK();
|
|
|
|
}
|
2018-12-31 20:04:05 +01:00
|
|
|
parser.skip('{');
|
|
|
|
if (parser.status().is_error()) {
|
2019-08-26 04:06:00 +02:00
|
|
|
return Status::Error(400, "Bad Request: JSON object expected");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
while (true) {
|
|
|
|
parser.skip_whitespaces();
|
|
|
|
if (parser.try_skip('}')) {
|
|
|
|
parser.skip_whitespaces();
|
|
|
|
if (parser.empty()) {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
return Status::Error(400, "Bad Request: unexpected data after object end");
|
|
|
|
}
|
|
|
|
if (parser.empty()) {
|
|
|
|
return Status::Error(400, "Bad Request: expected parameter name");
|
|
|
|
}
|
|
|
|
auto r_key = json_string_decode(parser);
|
|
|
|
if (r_key.is_error()) {
|
2018-07-08 01:47:46 +02:00
|
|
|
return Status::Error(400, PSLICE() << "Bad Request: can't parse parameter name: " << r_key.error().message());
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
parser.skip_whitespaces();
|
|
|
|
if (!parser.try_skip(':')) {
|
|
|
|
return Status::Error(400, "Bad Request: can't parse object, ':' expected");
|
|
|
|
}
|
|
|
|
parser.skip_whitespaces();
|
2018-06-27 18:24:42 +02:00
|
|
|
auto r_value = [&]() -> Result<MutableSlice> {
|
|
|
|
if (parser.peek_char() == '"') {
|
|
|
|
return json_string_decode(parser);
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
2018-06-27 18:24:42 +02:00
|
|
|
const int32 DEFAULT_MAX_DEPTH = 100;
|
|
|
|
auto begin = parser.ptr();
|
|
|
|
auto result = do_json_skip(parser, DEFAULT_MAX_DEPTH);
|
|
|
|
if (result.is_ok()) {
|
|
|
|
return MutableSlice(begin, parser.ptr());
|
|
|
|
} else {
|
|
|
|
return result.move_as_error();
|
|
|
|
}
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
2018-06-27 18:24:42 +02:00
|
|
|
}();
|
2018-12-31 20:04:05 +01:00
|
|
|
if (r_value.is_error()) {
|
2018-07-08 01:47:46 +02:00
|
|
|
return Status::Error(400, PSLICE() << "Bad Request: can't parse parameter value: " << r_value.error().message());
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
query_->args_.emplace_back(r_key.move_as_ok(), r_value.move_as_ok());
|
|
|
|
|
|
|
|
parser.skip_whitespaces();
|
|
|
|
if (parser.peek_char() != '}' && !parser.try_skip(',')) {
|
|
|
|
return Status::Error(400, "Bad Request: expected next field or object end");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
UNREACHABLE();
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2023-03-21 09:36:44 +01:00
|
|
|
Status HttpReader::parse_http_version(Slice version) {
|
|
|
|
if (version == "HTTP/1.1") {
|
|
|
|
query_->keep_alive_ = true;
|
|
|
|
} else if (version == "HTTP/1.0") {
|
|
|
|
query_->keep_alive_ = false;
|
|
|
|
} else {
|
|
|
|
LOG(INFO) << "Unsupported HTTP version: " << version;
|
|
|
|
return Status::Error(505, "HTTP Version Not Supported");
|
|
|
|
}
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2018-12-31 20:04:05 +01:00
|
|
|
Status HttpReader::parse_head(MutableSlice head) {
|
|
|
|
Parser parser(head);
|
|
|
|
|
|
|
|
Slice type = parser.read_till(' ');
|
|
|
|
parser.skip(' ');
|
|
|
|
// GET POST HTTP/1.1
|
|
|
|
if (type == "GET") {
|
2020-06-15 03:23:47 +02:00
|
|
|
query_->type_ = HttpQuery::Type::Get;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else if (type == "POST") {
|
2020-06-15 03:23:47 +02:00
|
|
|
query_->type_ = HttpQuery::Type::Post;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else if (type.size() >= 4 && type.substr(0, 4) == "HTTP") {
|
2023-03-21 09:36:44 +01:00
|
|
|
TRY_STATUS(parse_http_version(type));
|
|
|
|
query_->type_ = HttpQuery::Type::Response;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
|
|
|
LOG(INFO) << "Not Implemented " << tag("type", type) << tag("head", head);
|
|
|
|
return Status::Error(501, "Not Implemented");
|
|
|
|
}
|
|
|
|
|
|
|
|
query_->args_.clear();
|
|
|
|
|
2020-06-15 03:23:47 +02:00
|
|
|
if (query_->type_ == HttpQuery::Type::Response) {
|
2018-12-31 20:04:05 +01:00
|
|
|
query_->code_ = to_integer<int32>(parser.read_till(' '));
|
|
|
|
parser.skip(' ');
|
|
|
|
query_->reason_ = parser.read_till('\r');
|
2018-08-17 18:03:07 +02:00
|
|
|
LOG(DEBUG) << "Receive HTTP response " << query_->code_ << " " << query_->reason_;
|
2018-12-31 20:04:05 +01:00
|
|
|
} else {
|
|
|
|
auto url_version = parser.read_till('\r');
|
|
|
|
auto space_pos = url_version.rfind(' ');
|
|
|
|
if (space_pos == static_cast<size_t>(-1)) {
|
|
|
|
return Status::Error(400, "Bad Request: wrong request line");
|
|
|
|
}
|
|
|
|
|
|
|
|
TRY_STATUS(parse_url(url_version.substr(0, space_pos)));
|
2023-03-21 09:36:44 +01:00
|
|
|
TRY_STATUS(parse_http_version(url_version.substr(space_pos + 1)));
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
parser.skip('\r');
|
|
|
|
parser.skip('\n');
|
|
|
|
|
2023-11-25 00:45:04 +01:00
|
|
|
content_length_ = -1;
|
2019-02-13 00:29:52 +01:00
|
|
|
content_type_ = Slice("application/octet-stream");
|
2018-12-31 20:04:05 +01:00
|
|
|
content_type_lowercased_ = content_type_.str();
|
2019-02-13 00:29:52 +01:00
|
|
|
transfer_encoding_ = Slice();
|
|
|
|
content_encoding_ = Slice();
|
2018-12-31 20:04:05 +01:00
|
|
|
|
|
|
|
query_->headers_.clear();
|
|
|
|
query_->files_.clear();
|
|
|
|
query_->content_ = MutableSlice();
|
|
|
|
while (parser.status().is_ok() && !parser.data().empty()) {
|
|
|
|
MutableSlice header_name = parser.read_till(':');
|
|
|
|
parser.skip(':');
|
|
|
|
char *header_value_begin = parser.ptr();
|
|
|
|
char *header_value_end;
|
|
|
|
do {
|
|
|
|
parser.read_till('\r');
|
|
|
|
header_value_end = parser.ptr();
|
|
|
|
parser.skip('\r');
|
|
|
|
parser.skip('\n');
|
|
|
|
} while (parser.status().is_ok() && (parser.peek_char() == ' ' || parser.peek_char() == '\t'));
|
|
|
|
|
|
|
|
process_header(header_name, {header_value_begin, header_value_end});
|
|
|
|
}
|
|
|
|
return parser.status().is_ok() ? Status::OK() : Status::Error(400, "Bad Request");
|
|
|
|
}
|
|
|
|
|
|
|
|
Status HttpReader::open_temp_file(CSlice desired_file_name) {
|
|
|
|
CHECK(temp_file_.empty());
|
|
|
|
|
|
|
|
auto tmp_dir = get_temporary_dir();
|
|
|
|
if (tmp_dir.empty()) {
|
|
|
|
return Status::Error("Can't find temporary directory");
|
|
|
|
}
|
|
|
|
|
2018-01-26 14:47:46 +01:00
|
|
|
TRY_RESULT(dir, realpath(tmp_dir, true));
|
2018-12-31 20:04:05 +01:00
|
|
|
CHECK(!dir.empty());
|
2020-01-16 02:47:28 +01:00
|
|
|
|
2018-12-31 20:04:05 +01:00
|
|
|
auto first_try = try_open_temp_file(dir, desired_file_name);
|
|
|
|
if (first_try.is_ok()) {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
2020-01-16 02:47:28 +01:00
|
|
|
|
2018-12-31 20:04:05 +01:00
|
|
|
// Creation of new file with desired name has failed. Trying to create unique directory for it
|
|
|
|
TRY_RESULT(directory, mkdtemp(dir, TEMP_DIRECTORY_PREFIX));
|
|
|
|
auto second_try = try_open_temp_file(directory, desired_file_name);
|
|
|
|
if (second_try.is_ok()) {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
auto third_try = try_open_temp_file(directory, "file");
|
|
|
|
if (third_try.is_ok()) {
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2018-02-16 10:09:50 +01:00
|
|
|
rmdir(directory).ignore();
|
2023-04-16 20:09:51 +02:00
|
|
|
LOG(WARNING) << "Failed to create temporary file \"" << desired_file_name << "\": " << second_try.error();
|
2018-12-31 20:04:05 +01:00
|
|
|
return second_try.move_as_error();
|
|
|
|
}
|
|
|
|
|
|
|
|
Status HttpReader::try_open_temp_file(Slice directory_name, CSlice desired_file_name) {
|
|
|
|
CHECK(temp_file_.empty());
|
|
|
|
CHECK(!directory_name.empty());
|
|
|
|
|
|
|
|
string file_name = clean_filename(desired_file_name);
|
|
|
|
if (file_name.empty()) {
|
|
|
|
file_name = "file";
|
|
|
|
}
|
|
|
|
|
|
|
|
temp_file_name_.clear();
|
|
|
|
temp_file_name_.reserve(directory_name.size() + 1 + file_name.size());
|
|
|
|
temp_file_name_.append(directory_name.data(), directory_name.size());
|
|
|
|
if (temp_file_name_.back() != TD_DIR_SLASH) {
|
|
|
|
temp_file_name_ += TD_DIR_SLASH;
|
|
|
|
}
|
|
|
|
temp_file_name_.append(file_name.data(), file_name.size());
|
|
|
|
|
|
|
|
TRY_RESULT(opened_file, FileFd::open(temp_file_name_, FileFd::Write | FileFd::CreateNew, 0640));
|
|
|
|
|
|
|
|
file_size_ = 0;
|
|
|
|
temp_file_ = std::move(opened_file);
|
|
|
|
LOG(DEBUG) << "Created temporary file " << temp_file_name_;
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
|
|
|
Status HttpReader::save_file_part(BufferSlice &&file_part) {
|
|
|
|
file_size_ += narrow_cast<int64>(file_part.size());
|
|
|
|
if (file_size_ > MAX_FILE_SIZE) {
|
2018-08-27 14:52:47 +02:00
|
|
|
clean_temporary_file();
|
2018-12-31 20:04:05 +01:00
|
|
|
return Status::Error(
|
2018-08-27 14:52:47 +02:00
|
|
|
413, PSLICE() << "Request Entity Too Large: file of size " << file_size_ << " is too big to be uploaded");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
LOG(DEBUG) << "Save file part of size " << file_part.size() << " to file " << temp_file_name_;
|
|
|
|
auto result_written = temp_file_.write(file_part.as_slice());
|
|
|
|
if (result_written.is_error() || result_written.ok() != file_part.size()) {
|
2018-08-27 14:52:47 +02:00
|
|
|
clean_temporary_file();
|
2020-04-27 19:10:13 +02:00
|
|
|
return Status::Error(500, "Internal Server Error: can't upload the file");
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
return Status::OK();
|
|
|
|
}
|
|
|
|
|
2018-08-27 14:52:47 +02:00
|
|
|
void HttpReader::clean_temporary_file() {
|
|
|
|
string file_name = temp_file_name_;
|
|
|
|
close_temp_file();
|
|
|
|
delete_temp_file(file_name);
|
|
|
|
}
|
|
|
|
|
2018-12-31 20:04:05 +01:00
|
|
|
void HttpReader::close_temp_file() {
|
|
|
|
LOG(DEBUG) << "Close temporary file " << temp_file_name_;
|
|
|
|
CHECK(!temp_file_.empty());
|
|
|
|
temp_file_.close();
|
|
|
|
CHECK(temp_file_.empty());
|
|
|
|
temp_file_name_.clear();
|
|
|
|
}
|
|
|
|
|
|
|
|
void HttpReader::delete_temp_file(CSlice file_name) {
|
|
|
|
CHECK(!file_name.empty());
|
|
|
|
LOG(DEBUG) << "Unlink temporary file " << file_name;
|
|
|
|
unlink(file_name).ignore();
|
|
|
|
PathView path_view(file_name);
|
|
|
|
Slice parent = path_view.parent_dir();
|
|
|
|
const size_t prefix_length = std::strlen(TEMP_DIRECTORY_PREFIX);
|
|
|
|
if (parent.size() >= prefix_length + 7 &&
|
|
|
|
parent.substr(parent.size() - prefix_length - 7, prefix_length) == TEMP_DIRECTORY_PREFIX) {
|
|
|
|
LOG(DEBUG) << "Unlink temporary directory " << parent;
|
2018-07-08 01:47:46 +02:00
|
|
|
rmdir(PSLICE() << Slice(parent.data(), parent.size() - 1)).ignore();
|
2018-12-31 20:04:05 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
} // namespace td
|