Greg Hazel

I believe I have a fix for that here: https://github.com/CamHenlin/iMessageModule/issues/2

Unfortunately, https://github.com/libevent/libevent/pull/611 made this worse, since evhttp is willing to call `shutdown(evcon->fd, EVUTIL_SHUT_WR)` which disables writing of the underlying bev even if it would have been possible.

> > bufferevent_filter (in be_filter_unlink) calls bufferevent_free even if data is waiting in the underlying output buffer. > > You are calling `bufferevent_free(filter)` what you are expecting? Can you share...

That's possible, yes, but other things which take a `bufferevent` do not know they need to do that. (evhttp in this case).

Rebased. Yes, I only applied caching to the async resolver implemented in libevent. The blocking OS resolver implementations often have caching of their own.

And yes, on by default. It improves performance, and doesn't violate the TTL of the records.

Since just the client can accept inbound TCP connections, the only way to establish a vpn connection would be to set up a reverse tunnel. (Establishing a reverse connection over...

When trying to run udp2raw client on my server, I get `[FATAL]socket bind error`, since the UDP port my VPN is listening on is already open.

I also just ran in to this.

I found another potential solution; watch `IOHID` for Touch Bar events, and only enable `CGEventTap` for a brief period after. As long as the `CGEventTap` was created before the event,...