mirror of
https://github.com/yhirose/cpp-httplib
synced 2024-11-21 14:29:10 -07:00
Refactoring for reading content
This commit is contained in:
parent
3541fe8330
commit
fccb84f5e8
1 changed files with 143 additions and 138 deletions
281
httplib.h
281
httplib.h
|
@ -60,6 +60,7 @@ typedef int socket_t;
|
|||
#endif //_WIN32
|
||||
|
||||
#include <assert.h>
|
||||
#include <atomic>
|
||||
#include <fcntl.h>
|
||||
#include <fstream>
|
||||
#include <functional>
|
||||
|
@ -70,7 +71,6 @@ typedef int socket_t;
|
|||
#include <string>
|
||||
#include <sys/stat.h>
|
||||
#include <thread>
|
||||
#include <atomic>
|
||||
|
||||
#ifdef CPPHTTPLIB_OPENSSL_SUPPORT
|
||||
#include <openssl/err.h>
|
||||
|
@ -98,7 +98,7 @@ inline const unsigned char *ASN1_STRING_get0_data(const ASN1_STRING *asn1) {
|
|||
#define CPPHTTPLIB_READ_TIMEOUT_USECOND 0
|
||||
#define CPPHTTPLIB_REQUEST_URI_MAX_LENGTH 8192
|
||||
#define CPPHTTPLIB_PAYLOAD_MAX_LENGTH (std::numeric_limits<size_t>::max)()
|
||||
#define CPPHTTPLIB_RECV_BUFSIZ 4096
|
||||
#define CPPHTTPLIB_RECV_BUFSIZ size_t(4096u)
|
||||
|
||||
namespace httplib {
|
||||
|
||||
|
@ -445,6 +445,76 @@ private:
|
|||
*/
|
||||
namespace detail {
|
||||
|
||||
#ifdef CPPHTTPLIB_ZLIB_SUPPORT
|
||||
inline bool can_compress(const std::string &content_type) {
|
||||
return !content_type.find("text/") || content_type == "image/svg+xml" ||
|
||||
content_type == "application/javascript" ||
|
||||
content_type == "application/json" ||
|
||||
content_type == "application/xml" ||
|
||||
content_type == "application/xhtml+xml";
|
||||
}
|
||||
|
||||
inline void compress(std::string &content) {
|
||||
z_stream strm;
|
||||
strm.zalloc = Z_NULL;
|
||||
strm.zfree = Z_NULL;
|
||||
strm.opaque = Z_NULL;
|
||||
|
||||
auto ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION, Z_DEFLATED, 31, 8,
|
||||
Z_DEFAULT_STRATEGY);
|
||||
if (ret != Z_OK) { return; }
|
||||
|
||||
strm.avail_in = content.size();
|
||||
strm.next_in = (Bytef *)content.data();
|
||||
|
||||
std::string compressed;
|
||||
|
||||
const auto bufsiz = 16384;
|
||||
char buff[bufsiz];
|
||||
do {
|
||||
strm.avail_out = bufsiz;
|
||||
strm.next_out = (Bytef *)buff;
|
||||
deflate(&strm, Z_FINISH);
|
||||
compressed.append(buff, bufsiz - strm.avail_out);
|
||||
} while (strm.avail_out == 0);
|
||||
|
||||
content.swap(compressed);
|
||||
|
||||
deflateEnd(&strm);
|
||||
}
|
||||
|
||||
inline void decompress(std::string &content) {
|
||||
z_stream strm;
|
||||
strm.zalloc = Z_NULL;
|
||||
strm.zfree = Z_NULL;
|
||||
strm.opaque = Z_NULL;
|
||||
|
||||
// 15 is the value of wbits, which should be at the maximum possible value to
|
||||
// ensure that any gzip stream can be decoded. The offset of 16 specifies that
|
||||
// the stream to decompress will be formatted with a gzip wrapper.
|
||||
auto ret = inflateInit2(&strm, 16 + 15);
|
||||
if (ret != Z_OK) { return; }
|
||||
|
||||
strm.avail_in = content.size();
|
||||
strm.next_in = (Bytef *)content.data();
|
||||
|
||||
std::string decompressed;
|
||||
|
||||
const auto bufsiz = 16384;
|
||||
char buff[bufsiz];
|
||||
do {
|
||||
strm.avail_out = bufsiz;
|
||||
strm.next_out = (Bytef *)buff;
|
||||
inflate(&strm, Z_NO_FLUSH);
|
||||
decompressed.append(buff, bufsiz - strm.avail_out);
|
||||
} while (strm.avail_out == 0);
|
||||
|
||||
content.swap(decompressed);
|
||||
|
||||
inflateEnd(&strm);
|
||||
}
|
||||
#endif
|
||||
|
||||
template <class Fn> void split(const char *b, const char *e, char d, Fn fn) {
|
||||
int i = 0;
|
||||
int beg = 0;
|
||||
|
@ -869,14 +939,18 @@ inline bool read_headers(Stream &strm, Headers &headers) {
|
|||
return true;
|
||||
}
|
||||
|
||||
inline bool read_content_with_length(Stream &strm, std::string &out, size_t len,
|
||||
Progress progress) {
|
||||
out.assign(len, 0);
|
||||
template <typename T>
|
||||
inline bool read_content_with_length(Stream &strm, size_t len,
|
||||
Progress progress, T callback) {
|
||||
char buf[CPPHTTPLIB_RECV_BUFSIZ];
|
||||
|
||||
size_t r = 0;
|
||||
while (r < len) {
|
||||
auto n = strm.read(&out[r], len - r);
|
||||
auto n = strm.read(buf, std::min((len - r), CPPHTTPLIB_RECV_BUFSIZ));
|
||||
if (n <= 0) { return false; }
|
||||
|
||||
callback(buf, n);
|
||||
|
||||
r += n;
|
||||
|
||||
if (progress) {
|
||||
|
@ -891,13 +965,14 @@ inline void skip_content_with_length(Stream &strm, size_t len) {
|
|||
char buf[CPPHTTPLIB_RECV_BUFSIZ];
|
||||
size_t r = 0;
|
||||
while (r < len) {
|
||||
auto n = strm.read(buf, CPPHTTPLIB_RECV_BUFSIZ);
|
||||
auto n = strm.read(buf, std::min((len - r), CPPHTTPLIB_RECV_BUFSIZ));
|
||||
if (n <= 0) { return; }
|
||||
r += n;
|
||||
}
|
||||
}
|
||||
|
||||
inline bool read_content_without_length(Stream &strm, std::string &out) {
|
||||
template <typename T>
|
||||
inline bool read_content_without_length(Stream &strm, T callback) {
|
||||
char buf[CPPHTTPLIB_RECV_BUFSIZ];
|
||||
for (;;) {
|
||||
auto n = strm.read(buf, CPPHTTPLIB_RECV_BUFSIZ);
|
||||
|
@ -906,13 +981,14 @@ inline bool read_content_without_length(Stream &strm, std::string &out) {
|
|||
} else if (n == 0) {
|
||||
return true;
|
||||
}
|
||||
out.append(buf, n);
|
||||
callback(buf, n);
|
||||
}
|
||||
|
||||
return true;
|
||||
}
|
||||
|
||||
inline bool read_content_chunked(Stream &strm, std::string &out) {
|
||||
template <typename T>
|
||||
inline bool read_content_chunked(Stream &strm, T callback) {
|
||||
const auto bufsiz = 16;
|
||||
char buf[bufsiz];
|
||||
|
||||
|
@ -923,8 +999,7 @@ inline bool read_content_chunked(Stream &strm, std::string &out) {
|
|||
auto chunk_len = std::stoi(reader.ptr(), 0, 16);
|
||||
|
||||
while (chunk_len > 0) {
|
||||
std::string chunk;
|
||||
if (!read_content_with_length(strm, chunk, chunk_len, nullptr)) {
|
||||
if (!read_content_with_length(strm, chunk_len, nullptr, callback)) {
|
||||
return false;
|
||||
}
|
||||
|
||||
|
@ -932,8 +1007,6 @@ inline bool read_content_chunked(Stream &strm, std::string &out) {
|
|||
|
||||
if (strcmp(reader.ptr(), "\r\n")) { break; }
|
||||
|
||||
out += chunk;
|
||||
|
||||
if (!reader.getline()) { return false; }
|
||||
|
||||
chunk_len = std::stoi(reader.ptr(), 0, 16);
|
||||
|
@ -947,39 +1020,61 @@ inline bool read_content_chunked(Stream &strm, std::string &out) {
|
|||
return true;
|
||||
}
|
||||
|
||||
inline bool is_chunked_transfer_encoding(const Headers &headers) {
|
||||
return !strcasecmp(get_header_value(headers, "Transfer-Encoding", 0, ""),
|
||||
"chunked");
|
||||
}
|
||||
|
||||
template <typename T>
|
||||
bool read_content(Stream &strm, T &x, uint64_t payload_max_length,
|
||||
bool &exceed_payload_max_length,
|
||||
Progress progress = Progress()) {
|
||||
if (has_header(x.headers, "Content-Length")) {
|
||||
bool read_content(Stream &strm, T &x, uint64_t payload_max_length, int &status,
|
||||
Progress progress) {
|
||||
|
||||
#ifndef CPPHTTPLIB_ZLIB_SUPPORT
|
||||
if (x.get_header_value("Content-Encoding") == "gzip") {
|
||||
status = 415;
|
||||
return false;
|
||||
}
|
||||
#endif
|
||||
|
||||
auto callback = [&](const char *buf, size_t n) { x.body.append(buf, n); };
|
||||
|
||||
auto ret = true;
|
||||
auto exceed_payload_max_length = false;
|
||||
|
||||
if (is_chunked_transfer_encoding(x.headers)) {
|
||||
ret = read_content_chunked(strm, callback);
|
||||
} else if (!has_header(x.headers, "Content-Length")) {
|
||||
ret = read_content_without_length(strm, callback);
|
||||
} else {
|
||||
auto len = get_header_value_uint64(x.headers, "Content-Length", 0);
|
||||
if (len == 0) {
|
||||
const auto &encoding =
|
||||
get_header_value(x.headers, "Transfer-Encoding", 0, "");
|
||||
if (!strcasecmp(encoding, "chunked")) {
|
||||
return read_content_chunked(strm, x.body);
|
||||
if (len > 0) {
|
||||
if ((len > payload_max_length) ||
|
||||
// For 32-bit platform
|
||||
(sizeof(size_t) < sizeof(uint64_t) &&
|
||||
len > std::numeric_limits<size_t>::max())) {
|
||||
exceed_payload_max_length = true;
|
||||
skip_content_with_length(strm, len);
|
||||
ret = false;
|
||||
} else {
|
||||
// NOTE: We can remove it if it doesn't give us enough better
|
||||
// performance.
|
||||
x.body.reserve(len);
|
||||
ret = read_content_with_length(strm, len, progress, callback);
|
||||
}
|
||||
}
|
||||
|
||||
if ((len > payload_max_length) ||
|
||||
// For 32-bit platform
|
||||
(sizeof(size_t) < sizeof(uint64_t) &&
|
||||
len > std::numeric_limits<size_t>::max())) {
|
||||
exceed_payload_max_length = true;
|
||||
skip_content_with_length(strm, len);
|
||||
return false;
|
||||
}
|
||||
|
||||
return read_content_with_length(strm, x.body, len, progress);
|
||||
} else {
|
||||
const auto &encoding =
|
||||
get_header_value(x.headers, "Transfer-Encoding", 0, "");
|
||||
if (!strcasecmp(encoding, "chunked")) {
|
||||
return read_content_chunked(strm, x.body);
|
||||
}
|
||||
return read_content_without_length(strm, x.body);
|
||||
}
|
||||
return true;
|
||||
|
||||
if (ret) {
|
||||
#ifdef CPPHTTPLIB_ZLIB_SUPPORT
|
||||
if (x.get_header_value("Content-Encoding") == "gzip") {
|
||||
detail::decompress(x.body);
|
||||
}
|
||||
#endif
|
||||
} else {
|
||||
status = exceed_payload_max_length ? 413 : 400;
|
||||
}
|
||||
|
||||
return ret;
|
||||
}
|
||||
|
||||
template <typename T> inline void write_headers(Stream &strm, const T &info) {
|
||||
|
@ -1252,76 +1347,6 @@ inline void make_range_header_core(std::string &field, uint64_t value1,
|
|||
make_range_header_core(field, args...);
|
||||
}
|
||||
|
||||
#ifdef CPPHTTPLIB_ZLIB_SUPPORT
|
||||
inline bool can_compress(const std::string &content_type) {
|
||||
return !content_type.find("text/") || content_type == "image/svg+xml" ||
|
||||
content_type == "application/javascript" ||
|
||||
content_type == "application/json" ||
|
||||
content_type == "application/xml" ||
|
||||
content_type == "application/xhtml+xml";
|
||||
}
|
||||
|
||||
inline void compress(std::string &content) {
|
||||
z_stream strm;
|
||||
strm.zalloc = Z_NULL;
|
||||
strm.zfree = Z_NULL;
|
||||
strm.opaque = Z_NULL;
|
||||
|
||||
auto ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION, Z_DEFLATED, 31, 8,
|
||||
Z_DEFAULT_STRATEGY);
|
||||
if (ret != Z_OK) { return; }
|
||||
|
||||
strm.avail_in = content.size();
|
||||
strm.next_in = (Bytef *)content.data();
|
||||
|
||||
std::string compressed;
|
||||
|
||||
const auto bufsiz = 16384;
|
||||
char buff[bufsiz];
|
||||
do {
|
||||
strm.avail_out = bufsiz;
|
||||
strm.next_out = (Bytef *)buff;
|
||||
deflate(&strm, Z_FINISH);
|
||||
compressed.append(buff, bufsiz - strm.avail_out);
|
||||
} while (strm.avail_out == 0);
|
||||
|
||||
content.swap(compressed);
|
||||
|
||||
deflateEnd(&strm);
|
||||
}
|
||||
|
||||
inline void decompress(std::string &content) {
|
||||
z_stream strm;
|
||||
strm.zalloc = Z_NULL;
|
||||
strm.zfree = Z_NULL;
|
||||
strm.opaque = Z_NULL;
|
||||
|
||||
// 15 is the value of wbits, which should be at the maximum possible value to
|
||||
// ensure that any gzip stream can be decoded. The offset of 16 specifies that
|
||||
// the stream to decompress will be formatted with a gzip wrapper.
|
||||
auto ret = inflateInit2(&strm, 16 + 15);
|
||||
if (ret != Z_OK) { return; }
|
||||
|
||||
strm.avail_in = content.size();
|
||||
strm.next_in = (Bytef *)content.data();
|
||||
|
||||
std::string decompressed;
|
||||
|
||||
const auto bufsiz = 16384;
|
||||
char buff[bufsiz];
|
||||
do {
|
||||
strm.avail_out = bufsiz;
|
||||
strm.next_out = (Bytef *)buff;
|
||||
inflate(&strm, Z_NO_FLUSH);
|
||||
decompressed.append(buff, bufsiz - strm.avail_out);
|
||||
} while (strm.avail_out == 0);
|
||||
|
||||
content.swap(decompressed);
|
||||
|
||||
inflateEnd(&strm);
|
||||
}
|
||||
#endif
|
||||
|
||||
#ifdef _WIN32
|
||||
class WSInit {
|
||||
public:
|
||||
|
@ -1588,7 +1613,7 @@ inline bool Server::is_running() const { return is_running_; }
|
|||
inline void Server::stop() {
|
||||
if (is_running_) {
|
||||
assert(svr_sock_ != INVALID_SOCKET);
|
||||
std::atomic<socket_t> sock (svr_sock_.exchange(INVALID_SOCKET));
|
||||
std::atomic<socket_t> sock(svr_sock_.exchange(INVALID_SOCKET));
|
||||
detail::shutdown_socket(sock);
|
||||
detail::close_socket(sock);
|
||||
}
|
||||
|
@ -1878,26 +1903,14 @@ Server::process_request(Stream &strm, bool last_connection,
|
|||
|
||||
// Body
|
||||
if (req.method == "POST" || req.method == "PUT" || req.method == "PATCH") {
|
||||
bool exceed_payload_max_length = false;
|
||||
if (!detail::read_content(strm, req, payload_max_length_,
|
||||
exceed_payload_max_length)) {
|
||||
res.status = exceed_payload_max_length ? 413 : 400;
|
||||
if (!detail::read_content(strm, req, payload_max_length_, res.status,
|
||||
Progress())) {
|
||||
write_response(strm, last_connection, req, res);
|
||||
return !exceed_payload_max_length;
|
||||
return true;
|
||||
}
|
||||
|
||||
const auto &content_type = req.get_header_value("Content-Type");
|
||||
|
||||
if (req.get_header_value("Content-Encoding") == "gzip") {
|
||||
#ifdef CPPHTTPLIB_ZLIB_SUPPORT
|
||||
detail::decompress(req.body);
|
||||
#else
|
||||
res.status = 415;
|
||||
write_response(strm, last_connection, req, res);
|
||||
return true;
|
||||
#endif
|
||||
}
|
||||
|
||||
if (!content_type.find("application/x-www-form-urlencoded")) {
|
||||
detail::parse_query_text(req.body, req.params);
|
||||
} else if (!content_type.find("multipart/form-data")) {
|
||||
|
@ -2069,19 +2082,11 @@ inline bool Client::process_request(Stream &strm, Request &req, Response &res,
|
|||
|
||||
// Body
|
||||
if (req.method != "HEAD") {
|
||||
bool exceed_payload_max_length = false;
|
||||
int dummy_status;
|
||||
if (!detail::read_content(strm, res, std::numeric_limits<uint64_t>::max(),
|
||||
exceed_payload_max_length, req.progress)) {
|
||||
dummy_status, req.progress)) {
|
||||
return false;
|
||||
}
|
||||
|
||||
if (res.get_header_value("Content-Encoding") == "gzip") {
|
||||
#ifdef CPPHTTPLIB_ZLIB_SUPPORT
|
||||
detail::decompress(res.body);
|
||||
#else
|
||||
return false;
|
||||
#endif
|
||||
}
|
||||
}
|
||||
|
||||
return true;
|
||||
|
|
Loading…
Reference in a new issue