There are people who claim that they can listen while they talk and not lose a bit of information, but in practice that's not true, and either information stream is unreliable.

It makes perfect sense that wanting to write to a socket you don't pretend to read from it, it's not the task at hand. Does it really make sense to

• whenever a socket becomes writable, stop reading, and start writing, and at the same time
• whenever a socket becomes readable, stop writing, and start reading?

I have found my programs to be less confusing if they are driven by just one of these fileevents.

Generally, that means:

• in server code, 'readable'. When done with reading, process, start writing, and
• in client code, 'writable'. When done with writing, start reading, process.

--shmem

_($_=" "x(1<<5)."?\n".q·/)Oo.  G°\        /
                              /\_¯/(q    /
----------------------------  \__(m.====·.(_("always off the crowd"))."·
");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}