Re: Outstanding operations after TLS closure/renegotiation

[Hallvard B Furuseth <h.b.furuseth@usit.uio.no>]

Maybe this will be my last reply to myself...

I wrote:
> New clients -- I've just been trying to think of a clean way to write a
> client which makes use of a response with may or may not arrive.  In
> some cases one could of course just wait for a while and then exit with
> an error message.  But if one wants to try anything more complex, the
> code quickly gets ugly.

...like reestablishing TLS and retrying the operation (with a new
message ID) after a while:

> In particular handling error responses from the
> operations, which might e.g. have been attempted in the opposite order
> from how they were sent.

-- 
Hallvard