Re: [racket] ssl-tcp-unit.rkt restrictions

2010-09-06 Thread Neil Van Dyke
Thanks, Jay. I haven't looked thoroughly yet, but I suspect that simply moving the SSL client context object from a hidden variable to a public parameter would be a small change that's a lot more flexible. I'm working with the old PLT 4.2.5 "openssl" code early this week, so I might soon be

Re: [racket] ssl-tcp-unit.rkt restrictions

2010-09-06 Thread Jay McCarthy
I doubt there is a good reason for these restrictions. I imagine that it works this way because it is only used by a few people and this is all they needed. IIRC, only the hand-in server uses it in the tree. I'm happy to work with you to improve it, if I understood what exactly you wanted. Jay O

[racket] ssl-tcp-unit.rkt restrictions

2010-09-05 Thread Neil Van Dyke
Why does the 5.0.1 "ssl-tcp-unit.rkt" hide the SSL client context and require all the certificates and such to be provided at unit invocation(?) time? Or, an example practical question: how is one supposed to use SSL connection operations with different certificates available each time? (I d