[Vortex] Are content-types and transfer-encodings supposed to work?

[Francis Brosnan Blazquez]

Hi Jens,

> Since Vortex's messages don't begin with a CRLF, they are not valid  
> BEEP messages (parsing the MIME headers will fail.) This makes
> Vortex  
> unable to interoperate with any correct BEEP implementation. And
> it's  
> not possible to work around this, because a correct BEEP  
> implementation cannot  reliably parse such a message and tell
> whether  
> it's supposed to have headers or not (it can't tell the difference  
> between a message with a content-type, and one that just happens to  
> have a payload beginning "Content-Type:...".)
> 
> It's as though I had an HTTP library that sent requests like "GETT /  
> HTTP/1.1 ..."
> 
> The word "mutant" was sensationalist, but I am kind of shocked by  
> this, especially since it was apparently already known but not
> called  
> out anywhere. Maybe there aren't too many other BEEP implementations  
> or clients to be incompatible with, so I imagine most users are just  
> using vortex to talk to vortex, but situations like this aren't
> going  
> to improve the viability of BEEP as a standard.

Just to clarify; first, I'm agree about the "mutant" adjective, one
might understand it, but what about the "private" one?? I fully
disagree.

Second, BEEP it's an standard. I don't see the jump step between having
a bug in vortex and viability of BEEP as standard. One might accept BEEP
becomes stronger if current implementations are compromised to follow it
strictly, including fixing bug that could break backward compatibility;
but hey, that's exactly what we are doing...

Now, going to the point. The wrong MIME implementation is a bug, and, as
such, it will be fixed. Nothing more. 

There is nothing estrange about finding some protocol/software
implementation doing things in the wrong way. It's simple, this is
software; if a bug isn't hurting, but other ones do, where is going to
be focused fixing attention? In correctness or functionality?

The bad news would be that the project will be on doing the same without
fixing. This is not going to happen.

The issue here is to enable vortex engine to support already deployed
installation (this is key) and at the same time properly implement MIME,
which would be great for Vortex. Any comment, idea or suggestion about
this is welcome.

Finally, I would like to thank your help Jens, which is *really*
valuable, but I think you should moderate your appreciations to avoid
giving the impression that Vortex is totally broken! 

Cheers,

> —Jens
-- 
Francis Brosnan Blazquez <francis at aspl.es>