Re: Hessian PHP implementation

On Jan 13, 2005, at 9:17 AM, Christian Campo wrote:

I've filed these as bug reports. Now that Hessian is starting to take
off, it can become a higher priority.

> - better error handling (HessianRuntimeException or
> HessianProtocolException is most times not good enough)

True.

> - did you ever try to create a webservice that had two methods with
> the same name but different signature ??? My users tried that and all
> you get is "Unexpected code: M" :-) I know its forbidden, but a
> clearer message would
> help

That should be allowable. The kind of overloading you can do with
scripting languages is a little trickier.

There's two kinds of mangling that are (partially) supported by Resin's
servlet. First, a simple argument length mangling:

foo()
foo(a)
foo(b, c)

Resin's servlet (in AbstractSkeleton) will accept foo__0, foo__1, and
foo__2 for those different methods. Those should work fine for
scripting languages as well.

There's a second kind of mangling that's only supported by the EJB
Hessian stub currently. That mangling looks something like: foo__int,
foo__v, foo__int__int (I'd need to look up the mangling. I don't
think that's exactly right.)

But that second mangling wouldn't work for a scripting language without
some extra configuration like an IDL or some other annotation.

> - GZIP requests and responses, hopefully by using HTTP Headers so that
> the client triggers whether he wants GZIPping or not

I like the examine-the-first-byte method.

> - optionally logging error message on the server when they occur

This is an issue. Currently, Hessian should work on JDK 1.2 and 1.3.
If we add logging, it will require JDK 1.4. I don't know if that's a
problem or now.

> - and there is always more :-)
>
> how is that for a start ?

Sounds good to me.

-- Scott
> --
> christian campo (gmail.com)
Received on Fri 14 Jan 2005 09:36:23 -0800