David,

Thank you for that! I will give it a shot.

Phil

On Mon, Aug 1, 2016 at 6:27 AM, David Matthews <
david.matth...@prolingua.co.uk> wrote:

> Phil,
>
> fun socketToInt s =
>     SysWord.toInt(
>         Posix.FileSys.fdToWord(
>                 valOf(Posix.FileSys.iodToFD(Socket.ioDesc s))));
>
> val s: (Socket.active Socket.stream) INetSock.sock = INetSock.TCP.socket();
>
> socketToInt s;
>
> This will only work on Unix since the Posix structure isn't available on
> Windows.  Yes, it is quite long-winded.
>
> David
>
> On 30/07/2016 23:35, Phil Eaton wrote:
>
>> I've come to in impasse with Ponyo trying to access common REST APIs. I've
>> gotten this far ignoring SSL but it's time to start attacking that. Has
>> anyone done any work on SSL/TLS in Standard ML? I could not find anything
>> on SML/NJ and MLton email archives (and couldn't find a reasonable way to
>> search Poly/ML's).
>>
>> It seems like the Socket module makes the underlying socket int pointer
>> completely opaque. Since that pointer is what you'd need to make the tls
>> call to wrap it in a tls context, it seems like it won't be possible to
>> use
>> the Socket module. I hope I am mistaken because otherwise the necessary
>> step will be to reimplement the Socket module functionality in Ponyo.
>>
>> Any thoughts or suggestions on moving forward?
>>
>
> The problem is that the socket is not necessarily an int.  In Windows it's
> a UINT_PTR that is actually a Windows Handle.  That all gets hidden in the
> run-time system in Poly/ML and the value that is passed back to the ML code
> is a
> _______________________________________________
> polyml mailing list
> polyml@inf.ed.ac.uk
> http://lists.inf.ed.ac.uk/mailman/listinfo/polyml
>



-- 
Phil Eaton
_______________________________________________
polyml mailing list
polyml@inf.ed.ac.uk
http://lists.inf.ed.ac.uk/mailman/listinfo/polyml

Reply via email to