Re: [Gnash-dev] non-conformance to HTTP standards

[Eric Hughes]

At 04:46 AM 5/17/2007, Martin Guy wrote:
> > I've found what I consider something completely harebrained in the 
> > HTTP/1.1 standard, so I decided to violate it.
>
> That's a pretty bad principle to start from...

I didn't start there.  I ended up there.

> Forgive me, I'm having trouble understanding the issue.

It's entirely syntax.  It has nothing to do with either semantics (what it 
means) or pragmatics (what it does).

> Sorry, I don't understand what you mean. Are you saying that cygnal
> did act on these before, but now, in accordance with the standard,
> does not?

It didn't do much either way before.  I am writing a proper parser for HTTP 
protocols, which includes an HTTP Request line, which I am writing to 
accept a grammar larger than the spec calls for.  Strictly speaking, that's 
a standards violation.

This point deserve a clarification.  I'm not generating anything to 
standard.  I am accepting something not to standard, and therefore possibly 
masking a standards violation by some other piece of software.

> As far as I know, everything after the ? is server-side-software
> dependent and can be interpreted by a cgi script or what have you; is
> part of the cygnal spec that it responds to such URL extensions
> automatically?

I'm not addressing what any stuff after "?" means; that's a semantic 
issue.  I'm not address what Cygnal does with any stuff after "?"; that's a 
pragmatic issue.  I'm only concerned at present with the fact that a "?" is 
there; that's the syntactic issue.  The existence of a query (which starts 
with "?") itself is the point.  The standard is not written in such a way 
that you can plunk down a query after an otherwise-valid URI.  Go figure.

Eric