Re: pcp updates - yippee secure socket connections work

[Ken McDonell <kenj@xxxxxxxxxxxxxxxx>]

On 20/04/13 00:16, Frank Ch. Eigler wrote:
> ...
> Note that even the TIMEOUT_DEFAULT is a problem in that allows a
> naughty client to send one byte on channel, and lets pmcd sit there
> for a few seconds waiting for more, not servicing anyone else.  This
> is what a further rework inspired by the "continuation passing style"
> comment in pduread() could correct.
Point taken ... but that involves much more work as pduread() would need 
to have one non-blocking behaviour for pmcd (and mainloop changes to 
pmcd) and "classical" behaviour for everyone else as almost every 
application we build as part of pcp or pcp-gui uses this routine 
indirectly through __pmGetPDU().
Is someone keeping a wishlist / backlog of work items we agree should be 
done at some point in the future?