RE: Larger than 32bit signed?

A lot of other useful data types miss from the lib.
NULL for example would make it a breeze to move around stuff gotten out of databases.

But (imho), simplicity is one of the strengths of the spec, rather than a weakness: the core datatypes can be defined/used in pretty much any language/platform under the sun, and libs are a breeze to implement.

Otoh if you take the basic data types used in SOAP, which tend to be derived from XSD, there are so many that just deciding which one to pick for a particular variable can be a time-consuming and error-prone process (19 simple types plus those derived). And I guess many toolkits out there only have very limited support for many of those types.

If you use the same xmlrpc toolkit on both ends of the communication, you migth be lucky: the toolkit in question could possibly make use of the maximum native integer type available on the platform, and have no problem in receiving or sending 8-byte integers.
If the toolkit in question did not support 8-byte integers, yould could of course hack it into the toolkit.
Or just use a string value for that particular bigint data field.

If you are looking forward to real interop, then you are out of luck: the spec has been frozen since about 1999 and there are very little chances that a spec extension, whoever proposed it, would be implemented by a large part of the existing toolkits (I can count more than a dozen tolkits for php alone...).

Just my 2C
Gaetano Giunta

> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]]On Behalf
> Of Mark Ellzey
> Sent: Wednesday, February 08, 2006 3:32 PM
> To: [email protected]
> Subject: [xml-rpc] Larger than 32bit signed?
> 
> 
> I have recently taken up a project which entails output of statistical
> information. Many if not most of the other applications in the
> enterprise use XML-RPC as a communications method and I wanted to do
> the same for this one.
> 
> The problem that I run into is that the XML-RPC specification only
> states an "i4" or 4 bytes signed for integers. I find it hard to
> believe that no other developer has never wanted any number larger
> than the ones provided.
> 
> Has there been any research done in the area of adding unsigned or 64
> bit numbers to the specification? I understand I can just convert the
> number to a string and have the application have the logic to convert
> it, but that does not scale.
> 
> If this has been posted on the wrong mailing list I apologize 
> ahead of time.
> 
> Thank you,
> Mark
> 
> 
> 
> 
>  
> Yahoo! Groups Links
> 
> 
> 
>  
> 
> 
> 

 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/xml-rpc/

<*> To unsubscribe from this group, send an email to:
    [email protected]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/