Replace libevent with our own HTTP and socket-handling implementation

pinheadmz commented at 7:32 pm on March 13, 2025: member

This is a major component of removing libevent as a dependency of the project

It is based on #30988 but only in the sense that it copies the Sockman class introduced in that PR. The p2p / Connman refactor isn’t needed for HTTP and therefore this branch can be reviewed and merged independently of the p2p changes.

Commit strategy:

Import sockman.{h,cpp} from #30988
Assert current behavior of HTTP with additional functional tests, including copying from libevent’s tests
Implement a few helper functions for strings, timestamps, etc needed by HTTP protocol
Isolate the existing libevent-based HTTP server in a namespace http_libevent
Implement HTTP in a new namespace http_bitcoin (the namespace manages duplicate HTTPRequest definitions, etc)
Switch bitcoind from the libevent server to the new server
Clean up (delete http_libevent)

I am currently seeing about a 10% speed up in the functional tests on my own arm/macos machine.

Integration testing:

I am testing the new HTTP server by forking projects that integrate with bitcoin via HTTP and running their integration tests with bitcoind built from this branch (on Github actions). I will continue adding integrations over time, and re-running these CI tests as this branch gets rebased:

DrahtBot commented at 7:32 pm on March 13, 2025: contributor

The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.

Code Coverage & Benchmarks

For details see: https://corecheck.dev/bitcoin/bitcoin/pulls/32061.

Reviews

See the guideline for information on the review process.

Type	Reviewers
ACK	romanz
Concept ACK	laanwj, vasild, fjahr, w0xlt

If your review is incorrectly listed, please react with 👎 to this comment and the bot will ignore it on the next update.

Conflicts

Reviewers, this pull request conflicts with the following ones:

#32408 (tests: Expand HTTP coverage to assert libevent behavior by pinheadmz)
#31929 (http: Make server shutdown more robust by hodlinator)
#31672 (rpc: add cpu_load to getpeerinfo by vasild)
#27731 (Prevent file descriptor exhaustion from too many RPC calls by fjahr)
#26022 (Add util::ResultPtr class by ryanofsky)
#25665 (refactor: Add util::Result failure values, multiple error and warning messages by ryanofsky)

If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.

LLM Linter (✨ experimental)

Possible typos and grammar issues:

In httpserver.h comment: “Each connection is assigned an unique id of this type.” → “Each connection is assigned a unique id of this type.”
In SockMan::ConnectAndMakeId documentation: “…if the later is used…” → “…if the latter is used…”
No other typographic or grammatical errors were found.

pinheadmz marked this as a draft on Mar 13, 2025

DrahtBot added the label CI failed on Mar 13, 2025

DrahtBot commented at 8:29 pm on March 13, 2025: contributor

🚧 At least one of the CI tasks failed. Debug: https://github.com/bitcoin/bitcoin/runs/38735177073

Try to run the tests locally, according to the documentation. However, a CI failure may still happen due to a number of reasons, for example:

Possibly due to a silent merge conflict (the changes in this pull request being incompatible with the current code in the target branch). If so, make sure to rebase on the latest commit of the target branch.
A sanitizer issue, which can only be found by compiling with the sanitizer and running the affected test.
An intermittent issue.

Leave a comment here, if you need help tracking down a confusing failure.

laanwj added the label RPC/REST/ZMQ on Mar 14, 2025

laanwj commented at 12:52 pm on March 14, 2025: member

Concept ACK, nice work

vasild commented at 1:57 pm on March 14, 2025: contributor

Concept ACK

fjahr commented at 8:29 pm on March 14, 2025: contributor

Concept ACK

My understanding from the low-level networking discussion at CoreDev was that this wouldn’t build on top of sockman. I guess the devil is in the details but can you address that in what sense the current approach follows what was discussed there? Thanks!

pinheadmz commented at 0:03 am on March 15, 2025: member

My understanding from the low-level networking discussion at CoreDev was that this wouldn’t build on top of sockman. I guess the devil is in the details but can you address that in what sense the current approach follows what was discussed there? Thanks!

Sure, by coredev I had already written most of this implementation (based on sockman) but the performance was bad, and that was part of the motivation behind the deep-dive talk. However, by the end of the week I had reviewed that code in person with smart attendees and not only improved the performance of my code but started to improve performance vs master branch as well! Those updates came in the days just after the deep-dive discussion.

SOME kind of sockman is needed to replace libevent. The one @vasild wrote does actually seem to work well for this purpose as well as for p2p, and it would be “nice” to only have to maintain one I/O loop structure in bitcoind. @theuni is investigating how a sockman for http could be optimized if it had no other purpose, and I think that is the kind of feedback that will help us decide which path to take.

pinheadmz force-pushed on Mar 18, 2025

vasild commented at 12:22 pm on March 19, 2025: contributor

SOME kind of sockman is needed to replace libevent … it would be “nice” to only have to maintain one I/O loop structure in bitcoind.

:100:

pinheadmz force-pushed on Mar 19, 2025

DrahtBot added the label Needs rebase on Mar 20, 2025

pinheadmz force-pushed on Mar 20, 2025

DrahtBot removed the label Needs rebase on Mar 20, 2025

pinheadmz force-pushed on Mar 21, 2025

DrahtBot added the label Needs rebase on Mar 24, 2025

pinheadmz force-pushed on Mar 29, 2025

DrahtBot removed the label CI failed on Mar 30, 2025

DrahtBot removed the label Needs rebase on Mar 30, 2025

pinheadmz force-pushed on Mar 31, 2025

pinheadmz commented at 10:18 am on March 31, 2025: member

I rebased this branch on a single squashed commit from #30988 essentially just cherry-picking sockman.{h,cpp} by @vasild and leaving out the p2p refactor. This will make rebase maintenance on master a lot easier by reducing conflicting scope, and hopefully also makes review easier. It also means to some extent this PR can be merged independently of #30988, and also gives @theuni some room to rewrite a specific HTTP sockman if a more efficient purpose-focused module can be written. (Will update PR description in a moment)

I’ve finally gotten all CI to pass so I’m going to mark this PR as ready for review as I move on to integration testing with all the bitcoin client libraries I can find!

pinheadmz marked this as ready for review on Mar 31, 2025

pinheadmz renamed this:
~~[draft] Replace libevent with our own HTTP and socket-handling implementation~~
Replace libevent with our own HTTP and socket-handling implementation
on Mar 31, 2025

pinheadmz force-pushed on Apr 3, 2025

pinheadmz commented at 7:14 pm on April 3, 2025: member

rebase to a225633e34 includes a new test for “pipelining” HTTP requests (thanks @theuni for pointing out this oversight) and also adds a queue of requests to each HTTPClient to ensure requests are processed in series, in the order they were received.

Introduce a new low-level socket managing class `SockMan` cfe5eba446

test: cover http pipelining 00d18bbe09

test: cover -rpcservertimeout

Testing this requires adding an option to TestNode to force
the test framework to establish a new HTTP connection for
every RPC. Otherwise, attempting to reuse a persistent connection
would cause framework RPCs during startup and shutdown to fail.

5ec06529bb

test: cover "chunked" Transfer-Encoding f259121f30

string: implement ParseUInt64Hex 4c83785d1d

string: add CaseInsensitiveComparator

https://httpwg.org/specs/rfc9110.html#rfc.section.5.1
Field names in HTTP headers are case-insensitive. This
comparator will be used in the headers map to search by key.
In libevent these are compared in lowercase:
  evhttp_find_header()
  evutil_ascii_strcasecmp()
  EVUTIL_TOLOWER_()

144a777f86

pinheadmz force-pushed on Apr 3, 2025

DrahtBot added the label CI failed on Apr 3, 2025

DrahtBot removed the label CI failed on Apr 4, 2025

w0xlt commented at 7:06 pm on April 17, 2025: contributor

Concept ACK.

in src/common/sockman.cpp:396 in cfe5eba446 outdated

391+        LogPrintf("connection from %s dropped: non-selectable socket\n", them.ToStringAddrPort());
392+        return;
393+    }
394+
395+    // According to the internet TCP_NODELAY is not carried into accepted sockets
396+    // on all platforms.  Set it again here just to be sure.

romanz commented at 5:44 am on April 21, 2025:

nit: this comment seems to be incorrect, since TCP_NODELAY is only set once by SockMan.

pinheadmz commented at 1:53 pm on April 28, 2025:

This comment may need to be applied upstream in #30988

FWIW, the Sockman backend might not be what we end up going with, as @theuni and I are considering an alternative.

in src/util/time.cpp:127 in aeb8352a9d outdated

118@@ -116,6 +119,20 @@ std::optional<int64_t> ParseISO8601DateTime(std::string_view str)
119     return int64_t{TicksSinceEpoch<std::chrono::seconds>(tp)};
120 }
121 
122+std::string FormatRFC7231DateTime(int64_t nTime)
123+{
124+    const std::chrono::sys_seconds secs{std::chrono::seconds{nTime}};
125+    const auto days{std::chrono::floor<std::chrono::days>(secs)};
126+    // 1970-01-01 was a Thursday
127+    std::string weekday{weekdays[(days.time_since_epoch().count() + 4) % 7]};

romanz commented at 5:49 am on April 21, 2025:

nit: weekday and month can be a string_view.

pinheadmz commented at 5:38 pm on April 28, 2025:

done

in src/util/strencodings.cpp:503 in a1e151c774 outdated

496@@ -497,3 +497,10 @@ std::optional<uint64_t> ParseByteUnits(std::string_view str, ByteUnit default_mu
497     }
498     return *parsed_num * unit_amount;
499 }
500+
501+std::vector<std::byte> StringToBuffer(const std::string& str)
502+{
503+    return std::vector<std::byte>(

romanz commented at 6:03 am on April 21, 2025:

nit: maybe using std::span will be simpler? For example:

0    auto span = std::as_bytes(std::span(str));
1    return {span.begin(), span.end()};

pinheadmz commented at 3:38 pm on April 28, 2025:

Great thanks, done.

in src/util/string.cpp:30 in a1e151c774 outdated

25+
26+    auto line_start = it;
27+    std::string line{};
28+    while (it != end) {
29+        char c = static_cast<char>(*it);
30+        line += c;

romanz commented at 6:33 am on April 21, 2025:

Maybe it would be better (performance-wise) to create the string after the loop is over?

 0std::optional<std::string> LineReader::ReadLine()
 1{
 2    if (it == end) {
 3        return std::nullopt;
 4    }
 5
 6    auto line_start = it;
 7    size_t count = 0;
 8    while (it != end) {
 9        char c = static_cast<char>(*it);
10        ++it;
11        ++count;
12        if (c == '\n') break;
13        if (count >= max_read) throw std::runtime_error("max_read exceeded by LineReader");
14    }
15    const std::byte *data = &*line_start;
16    std::string line{reinterpret_cast<const char *>(data), count};
17    line = TrimString(line); // delete trailing \r and/or \n
18    return line;
19}

pinheadmz commented at 3:43 pm on April 28, 2025:

awesome, done

in src/util/string.cpp:36 in a1e151c774 outdated

31+        ++it;
32+        if (c == '\n') break;
33+        if ((size_t)std::distance(line_start, it) >= max_read) throw std::runtime_error("max_read exceeded by LineReader");
34+    }
35+
36+    line = TrimString(line); // delete trailing \r and/or \n

romanz commented at 6:35 am on April 21, 2025:

Can we return a string_view to the internal buffer? (preventing allocation & copy)

pinheadmz commented at 5:31 pm on April 28, 2025:

by internal buffer I assume you mean m_recv_buffer which im trying to copy from into the HTTPRequest object, then clear the buffer. I left another response about maybe changing the type of that buffer to reduce unnecessary copies, but am going to leave it alone for this current rebase. What I did here though was swap out TrimString() with TrimStringView() so at least that’s one less.

in src/util/string.cpp:45 in a1e151c774 outdated

40+// Ignores max_read but won't overflow
41+std::string LineReader::ReadLength(size_t len)
42+{
43+    if (len == 0) return "";
44+    if (Left() < len) throw std::runtime_error("Not enough data in buffer");
45+    std::string out(reinterpret_cast<const char*>(&(*it)), len);

romanz commented at 6:37 am on April 21, 2025:

Can we return a string_view to the internal buffer? (preventing allocation & copy)

pinheadmz commented at 5:36 pm on April 28, 2025:

I think I addressed this already by taking your suggestion here https://github.com/bitcoin/bitcoin/pull/32061/files#r2052024440

romanz commented at 10:15 am on May 3, 2025:

7d30118401 seems to still return a std::string:

0$ git grep ReadLength
1src/util/string.cpp:  std::string LineReader::ReadLength(size_t len)
2src/util/string.h:    std::string ReadLength(size_t len);

pinheadmz commented at 1:43 pm on May 9, 2025:

Oh I see I misunderstood. For now I want LineReader to return new string objects because it reads directly from the HTTPClient’s m_recv_buffer which is then erased. I know we can prevent that copy by being more clever with m_recv_buffer and I still plan on exploring that, maybe as a follow up? It’s nice and simple now and there is a clean separation between the socket layer receiving data and the HTTPRequest

romanz commented at 1:46 pm on May 9, 2025:

Sounds good, thanks for the context :)

in src/httpserver.cpp:787 in 70d003ca10 outdated

780@@ -781,3 +781,69 @@ void UnregisterHTTPHandler(const std::string &prefix, bool exactMatch)
781         pathHandlers.erase(i);
782     }
783 }
784+
785+
786+namespace http_bitcoin {
787+std::optional<std::string> HTTPHeaders::Find(const std::string key) const

romanz commented at 6:48 am on April 21, 2025:

nit: consider using string_view instead of string here and below.

pinheadmz commented at 7:42 pm on April 28, 2025:

done.

romanz commented at 10:23 am on May 3, 2025:

nit: key & value can also be a string_view (here and below)

pinheadmz commented at 1:48 pm on May 9, 2025:

Cool will address this nit on next rebase thanks!

in src/httpserver.cpp:840 in 70d003ca10 outdated

835+}
836+
837+std::string HTTPHeaders::Stringify() const
838+{
839+    std::string out;
840+    for (auto it = m_map.begin(); it != m_map.end(); ++it) {

romanz commented at 4:11 pm on April 21, 2025:

nit: maybe the following would be simpler?

0    for (const auto& [k, v] : m_map) {
1        out += k + ": " + v + "\r\n";
2    }

pinheadmz commented at 7:41 pm on April 28, 2025:

thanks I’ll take it!

in src/httpserver.cpp:643 in 482382bd14 outdated

915+    if (!req->LoadBody(reader)) return false;
916+
917+    // Remove the bytes read out of the buffer.
918+    // If one of the above calls throws an error, the caller must
919+    // catch it and disconnect the client.
920+    m_recv_buffer.erase(

romanz commented at 8:49 am on April 27, 2025:

nit: maybe it’s better to return a std::span from m_recv_buffer to avoid a copy here.

pinheadmz commented at 3:21 pm on April 28, 2025:

we’d still want to erase the data after fulfilling the request though (if I’m understanding your comment correctly). Would it make more sense if m_recv_buffer was a std::deque or something where we can truncate the structure without copying/moving/shifting left?

romanz commented at 5:53 am on April 30, 2025:

Would it make more sense if m_recv_buffer was a std::deque or something where we can truncate the structure without copying/moving/shifting left?

Sounds good :)

in src/httpserver.cpp:531 in 73c3c2e3d3 outdated

940@@ -940,6 +941,92 @@ bool HTTPRequest::LoadBody(LineReader& reader)
941     }
942 }
943 
944+void HTTPRequest::WriteReply(HTTPStatusCode status, std::span<const std::byte> reply_body)

romanz commented at 9:02 am on April 27, 2025:

Would it be possible to allow writing a reply composed of several span<byte>? It should allow sending while serializing a response object.

pinheadmz commented at 2:33 pm on April 28, 2025:

Could this be done in a followup? I actually just checked #30321 to see why we didn’t do this back then – I remember when that was authored and merged but I didn’t realize you were the author!

romanz commented at 5:54 am on April 30, 2025:

Sure, thanks :)

in src/httpserver.cpp:605 in 73c3c2e3d3 outdated

1009+        m_client->m_send_buffer.insert(m_client->m_send_buffer.end(), headers_bytes.begin(), headers_bytes.end());
1010+
1011+        // We've been using std::span up until now but it is finally time to copy
1012+        // data. The original data will go out of scope when WriteReply() returns.
1013+        // This is analogous to the memcpy() in libevent's evbuffer_add()
1014+        m_client->m_send_buffer.insert(m_client->m_send_buffer.end(), reply_body.begin(), reply_body.end());

romanz commented at 9:04 am on April 27, 2025:

Maybe use a list of buffers (instead a single FIFO buffer)?

pinheadmz commented at 3:06 pm on April 28, 2025:

I don’t understand – would that be an optimization? It’s not really a FIFO, since we insert at the end of the buffer here but send data from the begin() of the buffer in SendBytesFromBuffer()

romanz commented at 5:57 am on April 30, 2025:

I suggest considering a similar approach to libevent’s evbuffers:

evbuffers are represented using a linked list of memory chunks, with pointers to the first and last chunk in the chain.

https://libevent.org/doc/buffer_8h.html

romanz commented at 6:45 am on May 1, 2025:

Thinking about it, doing the copy here is definitely fine for now (we can optimize it in a following PR).

in src/httpserver.h:275 in 73c3c2e3d3 outdated

270+    // Response headers may be set in advance before response body is known
271+    HTTPHeaders m_response_headers;
272+    void WriteReply(HTTPStatusCode status, std::span<const std::byte> reply_body = {});
273+    void WriteReply(HTTPStatusCode status, const char* reply_body) {
274+        auto reply_body_view = std::string_view(reply_body);
275+        std::span<const std::byte> byte_span(reinterpret_cast<const std::byte*>(reply_body_view.data()), reply_body_view.size());

romanz commented at 9:16 am on April 27, 2025:

nit: can be simplified using

0    void WriteReply(HTTPStatusCode status, std::string_view reply_body_view)
1    {
2        WriteReply(status, std::as_bytes(std::span{reply_body_view}));
3    }

pinheadmz commented at 3:36 pm on April 28, 2025:

awesome thanks

in src/httpserver.h:137 in 6a6285d268 outdated

170+    {
171+        auto reply_body_view = std::string_view(reply_body);
172+        std::span<const std::byte> byte_span(reinterpret_cast<const std::byte*>(reply_body_view.data()), reply_body_view.size());
173+        WriteReply(status, byte_span);
174+    }
175+    void WriteReply(HTTPStatusCode status, const std::string& reply_body)

romanz commented at 9:20 am on April 27, 2025:

Can be removed if we’ll add WriteReply(HTTPStatusCode status, std::string_view reply_body_view) (see above).

pinheadmz commented at 3:36 pm on April 28, 2025:

👍

in src/httpserver.cpp:246 in 90761d5026 outdated

242@@ -243,18 +243,18 @@ static bool InitHTTPAllowList()
243 }
244 
245 /** HTTP request method as string - use for logging only */
246-std::string RequestMethodString(HTTPRequest::RequestMethod m)
247+std::string RequestMethodString(HTTPRequestMethod m)

romanz commented at 9:43 am on April 27, 2025:

nit:

0std::string_view RequestMethodString(HTTPRequestMethod m)

pinheadmz commented at 1:59 pm on April 28, 2025:

done

in src/httpserver.cpp:980 in b828fa1e29 outdated

975+    if (end == std::string::npos) {
976+        end = decoded_uri.length();
977+    }
978+    const std::string query{decoded_uri.substr(start + 1, end - start - 1)};
979+    // find requested parameter in query
980+    const std::vector<std::string> params{SplitString(query, "&")};

romanz commented at 9:44 am on April 27, 2025:

nit: wouldn’t it copy the substrings?

pinheadmz commented at 2:12 pm on April 28, 2025:

You’re right thanks, I haven’t been dilligent about using string_view when copying isn’t needed. Fixed that here and also in HTTPRequest::LoadControlData()

in src/httpserver.cpp:994 in b828fa1e29 outdated

989+        }
990+    }
991+    return std::nullopt;
992+}
993+
994+std::pair<bool, std::string> HTTPRequest::GetHeader(const std::string& hdr) const

romanz commented at 9:45 am on April 27, 2025:

nit:

0std::pair<bool, std::string_view> HTTPRequest::GetHeader(const std::string& hdr) const

pinheadmz commented at 2:28 pm on April 28, 2025:

thanks, for this fix I added a new commit instead of a fixup. The reason is because in the “story” of the commits, GetHeader() is already defined and used by the legacy HTTP server returning a string, and the goal at this point is to match the exiting API. But we can optimize it after the big “switch HTTP servers” commit.

romanz commented at 1:00 pm on April 27, 2025: contributor

LGTM, many thanks! Added some comments/suggestions above :)

time: implement and test RFC7231 timestamp string

HTTP 1.1 responses require a timestamp header with a
specific format, specified in:
https://www.rfc-editor.org/rfc/rfc7231#section-7.1.1.1

5c5de980b1

string: add LineReader

This is a helper struct to parse HTTP messages from data in buffers
from sockets. HTTP messages begin with headers which are
CRLF-terminated lines (\n or \r\n) followed by an arbitrary amount of
body data. Whitespace is trimmed from the field lines but not the body.

https://httpwg.org/specs/rfc9110.html#rfc.section.5

43b4d0d2f3

http: enclose libevent-dependent code in a namespace

This commit is a no-op to isolate HTTP methods and objects that
depend on libevent. Following commits will add replacement objects
and methods in a new namespace for testing and review before
switching over the server.

6dd5723178

http: Implement HTTPHeaders class

see:
https://www.rfc-editor.org/rfc/rfc2616#section-4.2
https://www.rfc-editor.org/rfc/rfc7231#section-5
https://www.rfc-editor.org/rfc/rfc7231#section-7
https://httpwg.org/specs/rfc9111.html#header.field.definitions

e95c6f5b65

http: Implement HTTPResponse class

HTTP Response message:
https://datatracker.ietf.org/doc/html/rfc1945#section-6

Status line (first line of response):
https://datatracker.ietf.org/doc/html/rfc1945#section-6.1

Status code definitions:
https://datatracker.ietf.org/doc/html/rfc1945#section-9

037a626499

http: Implement HTTPRequest class

HTTP Request message:
https://datatracker.ietf.org/doc/html/rfc1945#section-5

Request Line aka Control Line aka first line:
https://datatracker.ietf.org/doc/html/rfc1945#section-5.1

See message_read_status() in libevent http.c for how
`MORE_DATA_EXPECTED` is handled there

67aa660c77

http: Begin implementation of HTTPClient and HTTPServer df11fc1173

http: read requests from connected clients 980a9cd3d3

http: support "chunked" Transfer-Encoding d7778f426c

http: compose and send replies to connected clients b55725130a

http: disconnect clients ebdb6d446e

Allow http workers to send data optimistically as an optimization 59b54fef7c

http: use a queue to pipeline requests from each connected client

See https://www.rfc-editor.org/rfc/rfc7230#section-6.3.2

> A server MAY process a sequence of pipelined requests in
  parallel if they all have safe methods (Section 4.2.1 of [RFC7231]),
  but it MUST send the corresponding responses in the same order that
  the requests were received.

We choose NOT to process requests in parallel. They are executed in
the order recevied as well as responded to in the order received.
This prevents race conditions where old state may get sent in response
to requests that are very quick to process but were requested later on
in the queue.

4c28395074

define HTTP request methods at module level outside of class

This is a refactor to prepare for matching the API of HTTPRequest
definitions in both namespaces http_bitcoin and http_libevent. In
particular, to provide a consistent return type for GetRequestMethod()
in both classes.

7a1359064e

Add helper methods to HTTPRequest to match original API

These methods are called by http_request_cb() and are present in the
original http_libevent::HTTPRequest.

bbc9c55974

refactor: split http_request_cb into libevent callback and dispatch

The original function is passed to libevent as a callback when HTTP
requests are received and processed. It wrapped the libevent request
object in a http_libevent::HTTPRequest and then handed that off to
bitcoin for basic checks and finally dispatch to worker threads.

In this commit we split the function after the
http_libevent::HTTPRequest is created, and pass that object to a new
function that maintains the logic of checking and dispatching.

This will be the merge point for http_libevent and http_bitcoin,
where HTTPRequest objects from either namespace have the same
downstream lifecycle.

2bea6bc887

refactor: split HTTPBindAddresses into config parse and libevent setup

The original function was already naturally split into two chunks:
First, we parse and validate the users' RPC configuration for IPs and
ports. Next we bind libevent's http server to the appropriate
endpoints.

This commit splits these chunks into two separate functions, leaving
the argument parsing in the common space of the module and moving the
libevent-specific binding into the http_libevent namespace.

A future commit will implement http_bitcoin::HTTPBindAddresses to
bind the validate list of endpoints by the new HTTP server.

1ab4aac359

http: implement new server control methods to match legacy API 84e1818563

http: disconnect after idle timeout (-rpcservertimeout) b90f808e30

use CScheduler for HTTPRPCTimer

This removes the dependency on libevent for scheduled events,
like re-locking a wallet some time after decryption.

2950597694

http: switch servers from libevent to bitcoin be97857d14

fuzz: switch http_libevent::HTTPRequest to http_bitcoin::HTTPRequest 7856f20fec

http: using string_view to avoid unnecessary copy in Headers 54a36093f7

httpserver: delete libevent! 7d30118401

pinheadmz force-pushed on Apr 29, 2025

pinheadmz commented at 1:23 pm on April 29, 2025: member

Thanks so much for the great review @romanz I took most of your suggestions, there’s still an open question about m_recv_buffer but I really appreciate all your attention to string copying!

romanz commented at 7:54 am on May 3, 2025: contributor

tACK https://github.com/bitcoin/bitcoin/commit/7d301184016a3f59c2e363dff631263cdbe21da0

Tested with both electrs and bindex.

DrahtBot requested review from laanwj on May 3, 2025

DrahtBot requested review from fjahr on May 3, 2025

DrahtBot requested review from vasild on May 3, 2025

Replace libevent with our own HTTP and socket-handling implementation #32061

Integration testing:

Code Coverage & Benchmarks

Reviews

Conflicts

LLM Linter (✨ experimental)