bitcoin/src/httprpc.cpp

394 lines
15 KiB
C++
Raw Normal View History

// Copyright (c) 2015-2022 The Bitcoin Core developers
2015-01-05 21:40:24 +01:00
// Distributed under the MIT software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.
#include <httprpc.h>
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
#include <common/args.h>
#include <crypto/hmac_sha256.h>
#include <httpserver.h>
#include <logging.h>
#include <netaddress.h>
#include <rpc/protocol.h>
#include <rpc/server.h>
#include <util/fs.h>
#include <util/fs_helpers.h>
#include <util/strencodings.h>
#include <util/string.h>
#include <walletinitinterface.h>
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
2018-03-22 19:27:49 +01:00
#include <algorithm>
#include <iterator>
#include <map>
#include <memory>
#include <optional>
2018-03-22 19:27:49 +01:00
#include <set>
#include <string>
2022-04-30 13:05:30 +02:00
#include <vector>
using util::SplitString;
using util::TrimStringView;
/** WWW-Authenticate to present with 401 Unauthorized response */
static const char* WWW_AUTH_HEADER_DATA = "Basic realm=\"jsonrpc\"";
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
/** Simple one-shot callback timer to be used by the RPC mechanism to e.g.
2017-01-18 16:15:37 +01:00
* re-lock the wallet.
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
*/
class HTTPRPCTimer : public RPCTimerBase
{
public:
HTTPRPCTimer(struct event_base* eventBase, std::function<void()>& func, int64_t millis) :
ev(eventBase, false, func)
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
struct timeval tv;
tv.tv_sec = millis/1000;
tv.tv_usec = (millis%1000)*1000;
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
ev.trigger(&tv);
}
private:
HTTPEvent ev;
};
class HTTPRPCTimerInterface : public RPCTimerInterface
{
public:
explicit HTTPRPCTimerInterface(struct event_base* _base) : base(_base)
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
}
const char* Name() override
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
return "HTTP";
}
RPCTimerBase* NewTimer(std::function<void()>& func, int64_t millis) override
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
return new HTTPRPCTimer(base, func, millis);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
}
private:
struct event_base* base;
};
/* Pre-base64-encoded authentication token */
static std::string strRPCUserColonPass;
/* Stored RPC timer interface (for unregistration) */
static std::unique_ptr<HTTPRPCTimerInterface> httpRPCTimerInterface;
2020-11-17 00:53:36 +01:00
/* List of -rpcauth values */
static std::vector<std::vector<std::string>> g_rpcauth;
2018-03-22 19:27:49 +01:00
/* RPC Auth Whitelist */
static std::map<std::string, std::set<std::string>> g_rpc_whitelist;
static bool g_rpc_whitelist_default = false;
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
static void JSONErrorReply(HTTPRequest* req, UniValue objError, const JSONRPCRequest& jreq)
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
// Sending HTTP errors is a legacy JSON-RPC behavior.
Assume(jreq.m_json_version != JSONRPCVersion::V2);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
// Send error reply from json-rpc error object
int nStatus = HTTP_INTERNAL_SERVER_ERROR;
int code = objError.find_value("code").getInt<int>();
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
if (code == RPC_INVALID_REQUEST)
nStatus = HTTP_BAD_REQUEST;
else if (code == RPC_METHOD_NOT_FOUND)
nStatus = HTTP_NOT_FOUND;
std::string strReply = JSONRPCReplyObj(NullUniValue, std::move(objError), jreq.id, jreq.m_json_version).write() + "\n";
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
req->WriteHeader("Content-Type", "application/json");
req->WriteReply(nStatus, strReply);
}
//This function checks username and password against -rpcauth
//entries from config file.
static bool multiUserAuthorized(std::string strUserPass)
{
if (strUserPass.find(':') == std::string::npos) {
return false;
}
std::string strUser = strUserPass.substr(0, strUserPass.find(':'));
std::string strPass = strUserPass.substr(strUserPass.find(':') + 1);
2020-11-17 00:53:36 +01:00
for (const auto& vFields : g_rpcauth) {
std::string strName = vFields[0];
if (!TimingResistantEqual(strName, strUser)) {
continue;
}
std::string strSalt = vFields[1];
std::string strHash = vFields[2];
static const unsigned int KEY_SIZE = 32;
unsigned char out[KEY_SIZE];
CHMAC_SHA256(reinterpret_cast<const unsigned char*>(strSalt.data()), strSalt.size()).Write(reinterpret_cast<const unsigned char*>(strPass.data()), strPass.size()).Finalize(out);
std::vector<unsigned char> hexvec(out, out+KEY_SIZE);
std::string strHashFromPass = HexStr(hexvec);
if (TimingResistantEqual(strHashFromPass, strHash)) {
return true;
}
}
return false;
}
static bool RPCAuthorized(const std::string& strAuth, std::string& strAuthUsernameOut)
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
if (strRPCUserColonPass.empty()) // Belt-and-suspenders measure if InitRPCAuthentication was not called
return false;
if (strAuth.substr(0, 6) != "Basic ")
return false;
std::string_view strUserPass64 = TrimStringView(std::string_view{strAuth}.substr(6));
auto userpass_data = DecodeBase64(strUserPass64);
std::string strUserPass;
if (!userpass_data) return false;
strUserPass.assign(userpass_data->begin(), userpass_data->end());
if (strUserPass.find(':') != std::string::npos)
strAuthUsernameOut = strUserPass.substr(0, strUserPass.find(':'));
//Check if authorized under single-user field
if (TimingResistantEqual(strUserPass, strRPCUserColonPass)) {
return true;
}
return multiUserAuthorized(strUserPass);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
}
static bool HTTPReq_JSONRPC(const std::any& context, HTTPRequest* req)
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
// JSONRPC handles only POST
if (req->GetRequestMethod() != HTTPRequest::POST) {
req->WriteReply(HTTP_BAD_METHOD, "JSONRPC server handles only POST requests");
return false;
}
// Check authorization
std::pair<bool, std::string> authHeader = req->GetHeader("authorization");
if (!authHeader.first) {
req->WriteHeader("WWW-Authenticate", WWW_AUTH_HEADER_DATA);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
req->WriteReply(HTTP_UNAUTHORIZED);
return false;
}
JSONRPCRequest jreq;
jreq.context = context;
jreq.peerAddr = req->GetPeer().ToStringAddrPort();
if (!RPCAuthorized(authHeader.second, jreq.authUser)) {
LogPrintf("ThreadRPCServer incorrect password attempt from %s\n", jreq.peerAddr);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
/* Deter brute-forcing
If this results in a DoS the user really
shouldn't have their RPC port exposed. */
UninterruptibleSleep(std::chrono::milliseconds{250});
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
req->WriteHeader("WWW-Authenticate", WWW_AUTH_HEADER_DATA);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
req->WriteReply(HTTP_UNAUTHORIZED);
return false;
}
try {
// Parse request
UniValue valRequest;
if (!valRequest.read(req->ReadBody()))
throw JSONRPCError(RPC_PARSE_ERROR, "Parse error");
// Set the URI
jreq.URI = req->GetURI();
UniValue reply;
2018-03-22 19:27:49 +01:00
bool user_has_whitelist = g_rpc_whitelist.count(jreq.authUser);
if (!user_has_whitelist && g_rpc_whitelist_default) {
LogPrintf("RPC User %s not allowed to call any methods\n", jreq.authUser);
req->WriteReply(HTTP_FORBIDDEN);
return false;
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
// singleton request
2018-03-22 19:27:49 +01:00
} else if (valRequest.isObject()) {
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
jreq.parse(valRequest);
2018-03-22 19:27:49 +01:00
if (user_has_whitelist && !g_rpc_whitelist[jreq.authUser].count(jreq.strMethod)) {
LogPrintf("RPC User %s not allowed to call method %s\n", jreq.authUser, jreq.strMethod);
req->WriteReply(HTTP_FORBIDDEN);
return false;
}
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
// Legacy 1.0/1.1 behavior is for failed requests to throw
// exceptions which return HTTP errors and RPC errors to the client.
// 2.0 behavior is to catch exceptions and return HTTP success with
// RPC errors, as long as there is not an actual HTTP server error.
const bool catch_errors{jreq.m_json_version == JSONRPCVersion::V2};
reply = JSONRPCExec(jreq, catch_errors);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
if (jreq.IsNotification()) {
// Even though we do execute notifications, we do not respond to them
req->WriteReply(HTTP_NO_CONTENT);
return true;
}
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
// array of requests
2018-03-22 19:27:49 +01:00
} else if (valRequest.isArray()) {
// Check authorization for each request's method
2018-03-22 19:27:49 +01:00
if (user_has_whitelist) {
for (unsigned int reqIdx = 0; reqIdx < valRequest.size(); reqIdx++) {
if (!valRequest[reqIdx].isObject()) {
throw JSONRPCError(RPC_INVALID_REQUEST, "Invalid Request object");
} else {
const UniValue& request = valRequest[reqIdx].get_obj();
// Parse method
std::string strMethod = request.find_value("method").get_str();
2018-03-22 19:27:49 +01:00
if (!g_rpc_whitelist[jreq.authUser].count(strMethod)) {
LogPrintf("RPC User %s not allowed to call method %s\n", jreq.authUser, strMethod);
req->WriteReply(HTTP_FORBIDDEN);
return false;
}
}
}
}
// Execute each request
reply = UniValue::VARR;
for (size_t i{0}; i < valRequest.size(); ++i) {
// Batches never throw HTTP errors, they are always just included
// in "HTTP OK" responses. Notifications never get any response.
UniValue response;
try {
jreq.parse(valRequest[i]);
response = JSONRPCExec(jreq, /*catch_errors=*/true);
} catch (UniValue& e) {
response = JSONRPCReplyObj(NullUniValue, std::move(e), jreq.id, jreq.m_json_version);
} catch (const std::exception& e) {
response = JSONRPCReplyObj(NullUniValue, JSONRPCError(RPC_PARSE_ERROR, e.what()), jreq.id, jreq.m_json_version);
}
if (!jreq.IsNotification()) {
reply.push_back(std::move(response));
}
}
// Return no response for an all-notification batch, but only if the
// batch request is non-empty. Technically according to the JSON-RPC
// 2.0 spec, an empty batch request should also return no response,
// However, if the batch request is empty, it means the request did
// not contain any JSON-RPC version numbers, so returning an empty
// response could break backwards compatibility with old RPC clients
// relying on previous behavior. Return an empty array instead of an
// empty response in this case to favor being backwards compatible
// over complying with the JSON-RPC 2.0 spec in this case.
if (reply.size() == 0 && valRequest.size() > 0) {
req->WriteReply(HTTP_NO_CONTENT);
return true;
}
2018-03-22 19:27:49 +01:00
}
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
else
throw JSONRPCError(RPC_PARSE_ERROR, "Top-level object parse error");
req->WriteHeader("Content-Type", "application/json");
req->WriteReply(HTTP_OK, reply.write() + "\n");
} catch (UniValue& e) {
JSONErrorReply(req, std::move(e), jreq);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
return false;
} catch (const std::exception& e) {
JSONErrorReply(req, JSONRPCError(RPC_PARSE_ERROR, e.what()), jreq);
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
return false;
}
return true;
}
static bool InitRPCAuthentication()
{
if (gArgs.GetArg("-rpcpassword", "") == "")
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
LogInfo("Using random cookie authentication.\n");
std::optional<fs::perms> cookie_perms{std::nullopt};
auto cookie_perms_arg{gArgs.GetArg("-rpccookieperms")};
if (cookie_perms_arg) {
auto perm_opt = InterpretPermString(*cookie_perms_arg);
if (!perm_opt) {
LogInfo("Invalid -rpccookieperms=%s; must be one of 'owner', 'group', or 'all'.\n", *cookie_perms_arg);
return false;
}
cookie_perms = *perm_opt;
}
if (!GenerateAuthCookie(&strRPCUserColonPass, cookie_perms)) {
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
return false;
}
} else {
LogPrintf("Config options rpcuser and rpcpassword will soon be deprecated. Locally-run instances may remove rpcuser to use cookie-based auth, or may be replaced with rpcauth. Please see share/rpcauth for rpcauth auth generation.\n");
strRPCUserColonPass = gArgs.GetArg("-rpcuser", "") + ":" + gArgs.GetArg("-rpcpassword", "");
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
}
if (!gArgs.GetArgs("-rpcauth").empty()) {
LogInfo("Using rpcauth authentication.\n");
2020-11-23 13:00:50 +01:00
for (const std::string& rpcauth : gArgs.GetArgs("-rpcauth")) {
2022-04-30 13:05:30 +02:00
std::vector<std::string> fields{SplitString(rpcauth, ':')};
const std::vector<std::string> salt_hmac{SplitString(fields.back(), '$')};
if (fields.size() == 2 && salt_hmac.size() == 2) {
fields.pop_back();
fields.insert(fields.end(), salt_hmac.begin(), salt_hmac.end());
2020-11-17 00:53:36 +01:00
g_rpcauth.push_back(fields);
2020-11-23 13:00:50 +01:00
} else {
LogPrintf("Invalid -rpcauth argument.\n");
return false;
2020-11-17 00:53:36 +01:00
}
}
}
2018-03-22 19:27:49 +01:00
g_rpc_whitelist_default = gArgs.GetBoolArg("-rpcwhitelistdefault", gArgs.IsArgSet("-rpcwhitelist"));
for (const std::string& strRPCWhitelist : gArgs.GetArgs("-rpcwhitelist")) {
auto pos = strRPCWhitelist.find(':');
std::string strUser = strRPCWhitelist.substr(0, pos);
bool intersect = g_rpc_whitelist.count(strUser);
std::set<std::string>& whitelist = g_rpc_whitelist[strUser];
if (pos != std::string::npos) {
std::string strWhitelist = strRPCWhitelist.substr(pos + 1);
std::vector<std::string> whitelist_split = SplitString(strWhitelist, ", ");
std::set<std::string> new_whitelist{
std::make_move_iterator(whitelist_split.begin()),
std::make_move_iterator(whitelist_split.end())};
2018-03-22 19:27:49 +01:00
if (intersect) {
std::set<std::string> tmp_whitelist;
std::set_intersection(new_whitelist.begin(), new_whitelist.end(),
whitelist.begin(), whitelist.end(), std::inserter(tmp_whitelist, tmp_whitelist.end()));
new_whitelist = std::move(tmp_whitelist);
}
whitelist = std::move(new_whitelist);
}
}
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
return true;
}
bool StartHTTPRPC(const std::any& context)
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
{
LogDebug(BCLog::RPC, "Starting HTTP RPC server\n");
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
if (!InitRPCAuthentication())
return false;
auto handle_rpc = [context](HTTPRequest* req, const std::string&) { return HTTPReq_JSONRPC(context, req); };
RegisterHTTPHandler("/", true, handle_rpc);
if (g_wallet_init_interface.HasWalletSupport()) {
RegisterHTTPHandler("/wallet/", false, handle_rpc);
}
struct event_base* eventBase = EventBase();
assert(eventBase);
httpRPCTimerInterface = std::make_unique<HTTPRPCTimerInterface>(eventBase);
RPCSetTimerInterface(httpRPCTimerInterface.get());
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
return true;
}
void InterruptHTTPRPC()
{
LogDebug(BCLog::RPC, "Interrupting HTTP RPC server\n");
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
}
void StopHTTPRPC()
{
LogDebug(BCLog::RPC, "Stopping HTTP RPC server\n");
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
UnregisterHTTPHandler("/", true);
if (g_wallet_init_interface.HasWalletSupport()) {
UnregisterHTTPHandler("/wallet/", false);
}
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
if (httpRPCTimerInterface) {
RPCUnsetTimerInterface(httpRPCTimerInterface.get());
httpRPCTimerInterface.reset();
evhttpd implementation - *Replace usage of boost::asio with [libevent2](http://libevent.org/)*. boost::asio is not part of C++11, so unlike other boost there is no forwards-compatibility reason to stick with it. Together with #4738 (convert json_spirit to UniValue), this rids Bitcoin Core of the worst offenders with regard to compile-time slowness. - *Replace spit-and-duct-tape http server with evhttp*. Front-end http handling is handled by libevent, a work queue (with configurable depth and parallelism) is used to handle application requests. - *Wrap HTTP request in C++ class*; this makes the application code mostly HTTP-server-neutral - *Refactor RPC to move all http-specific code to a separate file*. Theoreticaly this can allow building without HTTP server but with another RPC backend, e.g. Qt's debug console (currently not implemented) or future RPC mechanisms people may want to use. - *HTTP dispatch mechanism*; services (e.g., RPC, REST) register which URL paths they want to handle. By using a proven, high-performance asynchronous networking library (also used by Tor) and HTTP server, problems such as #5674, #5655, #344 should be avoided. What works? bitcoind, bitcoin-cli, bitcoin-qt. Unit tests and RPC/REST tests pass. The aim for now is everything but SSL support. Configuration options: - `-rpcthreads`: repurposed as "number of work handler threads". Still defaults to 4. - `-rpcworkqueue`: maximum depth of work queue. When this is reached, new requests will return a 500 Internal Error. - `-rpctimeout`: inactivity time, in seconds, after which to disconnect a client. - `-debug=http`: low-level http activity logging
2015-01-23 07:53:17 +01:00
}
}