On 7/6/05, Mind Bridge <[EMAIL PROTECTED]> wrote:
> Hi Howard,
> 
> Any problems for me to implement adding sqeezer adaptors in the way adding
> ASOs are implemented (using a configration point)?

Nope, that's my "vision".

> 
> One other issue -- client persistence always uses the serialization approach
> at the moment. Is there a problem with using the registered data sqeezers?
> That would make URLs significantly shorter (and readable) in some cases. I
> understand that some quoting may be necessary, but I am wondering if there
> is a problem beyond that.

What's getting encoded is a series of PropertyChange objects, each
identifying the component id path (optional), the property name, and
the property value.  Each of these is streamed into an
ObjectOutputStream, the result is optionally gzipped, then mime
encoded.  So using the squeezer isn't the right approach.

I suppose you could have the query string be:

state:MyPage=Tfoo,bar,99,zip,,SFred

That would be:
 "T" prefix is "textual representation" (i.e., not binary)
 property foo of component bar is 99
 property zip (of the page) is the string "Fred"


In the existing code, the initial character identifies how the
remainder is "encoded".  Currently, its used to determine if a
GzipInputStream is needed.

I suspect you could change the code to generate THREE representations
(binary, binary gzip and textual), and keep the shortest encoding.

However, another thing I've wanted to look into was to add the ability
to pass the binary object stream through an encrypter/decrypter, for
security reasons. That would NOT mesh well with a textual
representation.


A final note; the serialization-based SqueezeAdaptor should share code
with the binary/binary gzipped approach and support optional
encryption as well.  It also doesn't need to use its mutant Base64
encoding; standard Base64 should work fine (since element attributes
are URL encoded anyway by the Markup Writer).


> 
> -mb
> 
> ----- Original Message -----
> From: "Howard Lewis Ship" <[EMAIL PROTECTED]>
> To: "Tapestry development" <[email protected]>
> Sent: Thursday, July 07, 2005 2:46 AM
> Subject: Re: Registering a squeezer adaptor
> 
> 
> Not really; I've been meaning to loop back, define a configuration
> point for setting adaptors, and add logic to the service impl to
> initialize from that configuration point.
> 
> On 7/5/05, Mind Bridge <[EMAIL PROTECTED]> wrote:
> > Hi,
> >
> > What is the best way to register a custom data squeezer adaptor in 4.0? Is
> > that functionality complete?
> >
> > -mb
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> >
> >
> 
> 
> --
> Howard M. Lewis Ship
> Independent J2EE / Open-Source Java Consultant
> Creator, Jakarta Tapestry
> Creator, Jakarta HiveMind
> 
> Professional Tapestry training, mentoring, support
> and project work.  http://howardlewisship.com
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 
> 
> 


-- 
Howard M. Lewis Ship
Independent J2EE / Open-Source Java Consultant
Creator, Jakarta Tapestry
Creator, Jakarta HiveMind

Professional Tapestry training, mentoring, support
and project work.  http://howardlewisship.com

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to