The BSDs let you pass path names longer than sizeof(s.sun_path), they read past the end when socklen > sizeof(struct sockaddr_un) up to a limit¹. Let's support that.

¹ I'm 80% sure it's 255 on all BSDs but it certainly is on Darwin and FreeBSD.

Currently we use a pretty naive removal of testing dependencies for SwiftPM by setting an environment variable. We should be able to use Rocket and update our Rakefile script to just remove testing dependencies on each release, and keep it in the development process thanks to that.

The idea came from this PR that I saw RxSwiftCommunity/RxOptional#83, which sounds reall

Regardless of what is decided in #2435, we should improve panic messages on context related panics.

This includes error messages regarding missing timer or IO drivers.

I believe that libevent has a dependency on OpenSSL (found this here). However, this isn't documented anywhere on the README and the lack of specifics here created a problem with my install on Windows 10. Steps to repro:

git clone https://github.com/libevent/libevent.git
mkdir libevent/.build
cd libevent/.build
cma

Issue description

In writing rust zmq bindings for tokio, the question came up about when zmq_send and zmq_recv might return EAGAIN when passed the ZMQ_DONTWAIT flag in a multipart message. The implication of the high water mark documentation is that it will only return EAGAIN on the first part of the multipart message, but I couldn't find it explicitly stated anywhere that this

ByteBufferView is a Collection of UInt8 which "views" into a ByteBuffer. Right now, we basically only implement the subscript which vends individual bytes. That's good enough to implement Collection/Sequence (and a bunch of more specialised sub-protocols) but it's not very fast at all.

Especially methods like firstIndex(of:) or firstIndex(where:) will be very slow because it'l

It would be useful to be able to toggle/flag into a process-level mode to group by pid and display more process-level information like the full command line.

• For the "Utilization by process name" window, we could group by pid rather than process name so that it's easier to find specific heavy network resource consumers.
• For the "Utilization by connection" window, including the pid + cmdlin

https://github.com/zerotier/ZeroTierOne/tree/master/controller
documentation says :--

**Dockerizing Controllers

ZeroTier network controllers can easily be run in Docker or other container systems. Since containers do not need to actually join networks, extra privilege options like "--device=/dev/net/tun --privileged" are not needed. You'll just need to map the local JSON API port of the run

Each connection will be filtered by one rule - the first match - it seems that the order of rules is based off the name of the rule.

So for example, if I had the following rules

• Block some advertising domain
• Allow port 80

If the port 80 rule had a name that sorted earlier than the blocking rule, then it would still permit http requests to the domain.

I can't be sure whether its the

The documentation for Object, Var and Varholder lead me to believe that given the following setup:

std::vector<DynamicStruct> testList;
  Object t1;
  t1.set("name","com.tl");
  t1.set("num1",false);
  Object j;
  j.set("testval",testList);

The following call would let me extract the values (similar to my other use of Object with primitives and basic strings)

` auto retB

This is more of an ongoing coding practice than an actual task. We should add code comments to document pointers passed to Windows ioctl and back.

Target: v1.0 release

To Reproduce

From the documentation Adding Parameters (body() in the docs should be body):

	Fuel.post("https://httpbin.org/post", listOf("foo" to "foo", "bar" to "bar"))
		.also { println(it.url) }
		.also { println(String(it.body.toByteArray())) }

Expected output

https://httpbin.org/post
"

Currently in our docs for open_signal_receiver, we have an example of catching SIGHUP and using it to reload the application configuration. While talking to @wgwz today re: python-trio/snekomatic#41, I discovered that this example is maybe not so great, because readers who aren't already steeped in Unix tr

Knative intends to add as little overhead to pure HTTP as possible. As such, some of our components are exposed to the hot-path of just about each HTTP request coming

I think we should re write the comment in English because I found that there are some foreigners followed this repo.

And if we write the comments in English, I believe that there are more and more international developers get involved in this.

It would be handy (especially for intellisense purposes) to have XMLdoc equivalents for the existing Javadoc comments that have been copied from the Netty project.

If anyone wants to help out with this, please let me know and we can coordinate (maybe we can split up the work by namespace or something).

Examples in the documentation suggest that // can be used for comments. E.g. the CNI spec
contains an example like

{
  "cniVersion": "0.4.0",
  "name": "dbnet",
  "type": "bridge",
  // type (plugin) specific
  "bridge": "cni0",
  ...

Unfortunately, using such a configuration with `po

Some counters that would be useful:

• one that increments on every jit.flush(), whether provoked internally or externally
• one that increments on every stop-the-world GC (potentially more here, e.g. minor GC etc)

I know we'll have an auditlog with raptorjit but counters are useful too!