RE: ! Pling data type - Are you 'to spec' ???

["Edward Pearson" <edward.mailgroup@xxxxxxxxxxxxxxxx>]

------=_NextPart_000_0033_01C546AC.B495C2C0
Content-Type: text/plain;
charset="US-ASCII"
Content-Transfer-Encoding: 7bit

Right, I see now.

NO, NO, NO!

Use of the '!' must be explicitly stated in the schema definition.
Parameter/value pairs should be defined to use either '=' or '!' but not
both.

We also need to get straight on the use of alternative character encodings
since these will lead to incompatibilities.

_____

From: xAP_developer@xxxxxxx [mailto:xAP_developer@xxxxxxx]
On Behalf Of Kevin Hawkins
Sent: 21 April 2005 17:59
To: xAP_developer@xxxxxxx
Subject: Re: [xAP_developer] ! Pling data type - Are you 'to spec' ???


Edward Pearson wrote:

>I understand that the spec allows schema designers to choose binary
format
>for appropriate fields but you seem to be suggesting
>
errr not at all -   it was a question..

Kevin Hawkins said ...

>What I'm really asking is should a xAP compliant listener transparently
handle either coding on
>any parameter.


>("in BSC is it
>allowable, indeed
>you must be able, to send State!4F4646 as well as State=OFF") that
there is
>equivalence between the formats and that all applications should expect
any
>field to be in either format and to be able to deal with both formats
>dynamically. That's way off my interpretation of the spec, will break
>everything and is very hard to implement in things like Perl that work
>mostly on pattern matching. Please clarify!
>
>
As I said above it's a question and one that I too am hoping the answer
to is 'no' as in my closing comment..

Kevin Hawkins said

>(hoping its permissable that the schema defines the type as ASCII
> only - not binary)



The real point to the post was that people mustn't rely on the '=' sign
being present as it could be a ! character in binary data.  Anyone just
looking for the first = after a chr(10) would fail to recover
parameters/values correctly.

Kevin

>________________________________
>
>From: xAP_developer@xxxxxxx [mailto:xAP_developer@xxxxxxx]
>On Behalf Of Kevin Hawkins
>Sent: 21 April 2005 15:37
>To: xAP_developer@xxxxxxx
>Subject: [xAP_developer] ! Pling data type - Are you 'to spec' ???
>
>
>The xAP spec prohibits codes outside normal ASCII printable characters
>being used as parameter values and offers the use of the pling (!)
>instead to handle the full range of 8 bit values
>
>value=A
>value!41
>
>being equivalent.     Now I know very little has been implemented using
>pling but all is about to change with the release of xAP Netiom (early
>May) which necessarily has to use pling to handle transparent
>send/receive data over its serial port.  Phaedrus are also using the
xAP
>HomeSeer plugin to achieve plug and play via BSC.  Now I know Stuart
has
>made the necessary changes in xFX and James has been a beta tester for
>the xAP Netiom so I'm hoping we've covered the main angles here.
>
>BTW In Netiom currently you can configure it to use either and if using
>= it will strip any data from the data stream that is outside the
>permissable range.
>
>So just a 'heads up' to everyone else really to ensure that your code
>does conform to the xAP spec and handle ! correctly.  ( Not sue if mine
>does or doesn't - will have to play - does the OCX handle it
>transparently Patrick ? )
>
>One interesting thing  in the above examples.... should a script engine
>handle the values as the same and would it be permissable for example
to
>resend something received as =A  as !41 transparently or should a value
>stay within the expected =/!  encoding.  What I'm really asking  is
>should a xAP compliant listener transparently handle either coding on
>any parameter.     Leading on from that in BSC is it allowable (indeed
>you must be able) to send
>
>State!4F4646  as well as State=OFF
>
>if so I've a lot of work to do :-(
>
>Kevin (hoping its permissable that the schema defines the type as ASCII
>only - not binary)
>
>
>
>



_____