RE: [wsrp] I think we need an errata on perUser

["Nader Oteifa" <nader2@netunitysoftware.com>]

We scope in the following way:

 

consumer session ---> producer session --> registration session --> group session-->portlet session

 

Some of these session are purely for our purposes (e.g. caching).

 

Ultimately, all active portlets belong to some group session and the group session indicates if initCookie has been called or not.  If protocol is none, then a group session is created which contains all the active portlets and is marked as having already called initCookie even though the call was not really made.  If perUser, then a single group session is created that contains all the portlets and initCookie flag is set appropriately.  The end result is everything appears to be in some group session and initCookie is called when a new group session is created unless the protocol is none. 

 

Whether your consumer creates a new session when the user logs in or promotes the anonymous session to an authenticated session is consumer implementation dependent. I don’t see anything about it in WSRP.   Currently, we always create a new session when the user login changes even from anonymous. 

 

Nader

From: Michael Freedman [mailto:michael.freedman@oracle.com]
Sent: Wednesday, January 27, 2010 4:04 PM
To: Nathan Lipke
Cc: WSRP; Tim Lake
Subject: Re: [wsrp] I think we need an errata on perUser

 

I believe its (1).

On 1/21/2010 4:20 PM, Nathan Lipke wrote:

We have some confusion on the interpretation of perUser:

1. perUser is a misnomer and should be: per-consumer-session

1. initCookie() is called the 1st time the consumer makes a markup operation (as anonymous)
2. A change of user (e.g. login) does not necessitate a new call to initCookie()

2. perUser means really means perUser

1. initCookie() the 1st time the consumer makes a markup operation (as anonymous)
2. initCookie() is called again when the user changes (e.g. login) (as the logged in user)

 

Which is the intent, I asked for clarity on this during public review:

http://wiki.oasis-open.org/wsrp/WSRP_v2_Public_Review_comments

The resolution was to leave the wording as is.

 

Nate