Ok, thanks. So are you saying that the gadget itself should be able to
use a generic appData key ("data1") and the container will guarantee
that the personId/key is uniqueified with the appUrl/pagePosition based
on the passed security token?

The out-of-the-box Shindig just directly uses the userId based on
"UserId.getUserId(SecurityToken)" which only uses the SecurityToken's
OwnerId/ViewerId (it ignores the appId/appUrl). So it looks like all
gadgets for the same viewer share the same "persistent space" by
default, correct?

So it looks like in order to get it working correctly
"UserId.getUserId(SecurityToken)" cannot be used and custom code must be
written to create a userId based on the rest of the fields in the
SecurityToken, correct (or by the passed appId in each method)?

Thanks,
Dave


-----Original Message-----
From: Henning P. Schmiedehausen [mailto:[EMAIL PROTECTED] 
Sent: Friday, December 05, 2008 3:38 PM
To: [email protected]
Subject: Re: Scoping "newUpdatePersonAppDataRequest" persistence per
gadget?

"Citron, David" <[EMAIL PROTECTED]> writes:

>Question: there appears to be nothing in
"newUpdatePersonAppDataRequest"
>to scope the request to the current gadget (only to, say, the viewer).
>And the default Shindig AppDataService doesn't seem to.

>However, the iGoogle developer sandbox definitely seems to have
>persistent data isolation from gadget to gadget.

>How is this done? And can a gadget assume that its persistent data is
>visible only to itself or does it need to include some hopefully-unique
>gadget-specific substring in the data key?

Do you mean "if you have the same gadget on the same page twice, they
have different appdata" or "if you have two different gadgets on the
same page, they have different appdata"?

In the first case, you use an internal reference which maps gadget to
page position (that should be different for every gadget installed),
in the second, you use the appUrl (maybe through a hash) as a
distinction. We e.g. keep that information in the secure token that
our endpoints can decrypt.

    Ciao
        Henning

-- 
Henning P. Schmiedehausen - Palo Alto, California, U.S.A.
[EMAIL PROTECTED] "We're Germans and we use Unix. 
[EMAIL PROTECTED]          That's a combination of two demographic
groups
                            known to have no sense of humour
whatsoever."
                               -- Hanno Mueller,
de.comp.os.unix.programming 

Reply via email to