cocoon-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nicola Ken Barozzi" <nicola...@supereva.it>
Subject Re: Dynamic XML generation
Date Wed, 19 Jul 2000 17:39:28 GMT
----- Original Message ----- 
From: "Ulrich Mayring" <ulim@denic.de>

> Jeremy Quinn wrote:
> > 
> > I am assuming (I never used session before) that it would be possible for
> > both of our taglibs to only create a new session if one did not already
> > exist, and then to keep "personal" values in there, like FP puts things
> > like fp-action="blah" into parent tags, hoping that the user of the tagset,
> > or another tagset is not going to use the same name.
> > 
> > FP would make session objects (eg. Hashtable) named "fp.vars" or something,
> > you use your naming scheme. We each prefix by the name of our own tagset
> > and Bob's yer uncle, no trampling.
> 
> Well, suppose I create a session, because the user successfully
> authenticated. I put some values in it, then your fp taglib puts some
> more stuff in it. Then the user leaves the confidential area and I kill
> the session and your values are gone. The same thing happens if we do it
> the other way round: you create a session, because the user is beginning
> to use fp. Then he accesses confidential areas and needs to
> authenticate. So I grab your session and put my values in it. Later on
> the user stops using fp, so you kill the session and the user is also
> dis-authenticated automatically.
> 
> I think cooperation can only work, if it is 100% clear who creates and
> kills sessions. This calls for a seperate taglib that does just that.
> 

Maybe session management should be encapsulated like the output
stream is?
It's not possible that every taglib author has to agree with other authors...
Maybe not making it possible to close sessions?
Hmm...

nicola_ken


Mime
View raw message