Unfortunately, to access this new field from my gadget token I'm gonna need
Stanton's latest fix, unfortunately this means reintroducing the metadata
bug I reported.  CATCH-22. Argh!  Hopefully both can be fixed.  Still
haven't tracked down the metadata issue other than the osapi change made
last week.

doug


On 9/21/11 2:19 PM, "Henry Saputra" <[email protected]> wrote:

> Ah cool, I am not aware about the field. Thanks Stanton.
> 
> - Henry
> 
> On Wed, Sep 21, 2011 at 7:26 AM, Stanton Sievers <[email protected]> wrote:
>> Hi Henry and Doug,
>> 
>> Is this not the purpose of SecurityToken.getTrustedJson()? ?It seems to
>> just be some arbitrary value you can stick in the token. ?If this is not
>> the purpose of the trusted json, do you know what is?
>> 
>> Thanks,
>> -Stanton
>> 
>> 
>> 
>> From: ? Henry Saputra <[email protected]>
>> To: ? ? [email protected],
>> Date: ? 09/21/2011 10:03
>> Subject: ? ? ? ?Re: Adding fields to security context v.s. overloading
>> viewerId
>> 
>> 
>> 
>> Hi Doug,
>> 
>> We have our own security token but havent moved to common container or
>> use the new Gadgets metadata APIs so didnt have this problem ... yet
>> =)
>> 
>> I am thinking about adding a common property to the SecurityToken
>> interface itself to support fields extension like this.
>> 
>> Let me know if you want to take a stab at it otherwise I could propose
>> something later today.
>> 
>> - Henry
>> 
>> On Wed, Sep 21, 2011 at 6:38 AM, daviesd <[email protected]> wrote:
>>> Yesterday I sent a message to the listserv describing what we have done
>> with
>>> our SecurityToken implementation. ?We¹ve extended SecurityToken and
>> created
>>> OurBlobCrypterSecurityToken. ?It has 1 new field that is a complex
>> structure
>>> (stored as json). ?This was not added to the SecurityToken interface,
>> but
>>> only to our class and when we need access to these values I was casting
>> the
>>> object.
>>> 
>>> With the recent issue in encodeToken with it using the proxy, I don¹t
>> have
>>> access to copy this over during the gadget iframe security token
>> generation.
>>> In fact, I¹m not even sure at the higher level (
>>> 
>> org.apache.shindig.gadgets.servlet.GadgetsHandlerService.convertAuthContext
>>> ) that it will copy the container security token fields over to the
>> gadget
>>> one correctly with this new field.
>>> 
>>> So my question is, how are people extending the security token? ?This
>> new
>>> field I added is additional info we need (security privileges and some
>> group
>>> information) that is associated with the viewerId. ?Should I really just
>>> overload viewerId with this json object and then whenever I use viewerId
>> in
>>> my services I would pull the appropriate values out of the json? ?Or is
>> the
>>> security token really meant to be extended?
>>> 
>>> Help is appreciated.
>>> 
>>> Doug
>>> 
>>> 
>> 
>> 
>> 
>> 
> 


Reply via email to