jakarta-cactus-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Pedro Nevado" <pnev...@javato.com>
Subject RE: FilterRedirector does not initialize sessions. - IS THIS A BUG ??
Date Thu, 11 Nov 2004 17:39:16 GMT
Hi there,

As far as I remember, the problem is not only with the default mecanism to
associate a session to the request in the client side. Even if you
explicitly set automatic session to true in the client side, this statement
is ineffective in the server side.

In my test with filters I bypass this limitation dealing with the session in
the server side.

A midway would be just to patch the code to allow instatiating a session if
explicitly said in the webrequest. That way, there would not be, I think,
any effect on existing tests wich counts on the present default mecanism.

Please, let me now what you finally conclude about this point.
Regards,

Pedro

-----Mensaje original-----
De: Kazuhito SUGURI [mailto:suguri.kazuhito@lab.ntt.co.jp]
Enviado el: jueves, 11 de noviembre de 2004 17:14
Para: cactus-user@jakarta.apache.org
Asunto: Re: FilterRedirector does not initialize sessions. - IS THIS A
BUG ??


Hi Vincent,

In article <200411111303.iABD3D8r016548@tama5.ecl.ntt.co.jp>,
Thu, 11 Nov 2004 14:03:04 +0100,
"Vincent Massol" <vmassol@pivolis.com> wrote:
vmassol> What strategy would you like us to follow to align the behaviors of
vmassol> FilterTestCase and ServletTestCase?

To align...
Is this means that FilterTestCase and ServletTestCase should
have same behavior of its default setAutomateSession()?
If so, I think it's impossible without breaking existing TestCases.
Of cource, it might be more trifling problem than I think.

Why don't we wait for reactions from other users of this list?
# How do you think, Pedro?


vmassol> BTW, do you have any idea why I decided to enable session creation
by
vmassol> default for ServletTestCase? I can't recall... ;-)

I'm not sure since it had already been so when I used Cactus
for the first time. But I can guess some:
1. Servlet uses a session object in many cases, then, it's user friendly
   if the session object is prepared by default.
2. if method under test uses a session object, a TestCase
   'there is no session object available' is needed only once,
   however, there are other many 'when the session object has XYZ'
TestCases.
3. if method under test does not use any session object,
   the default behavior of the session creation can not be a problem.


Regards,
----
Kazuhito SUGURI
mailto:suguri.kazuhito@lab.ntt.co.jp

---------------------------------------------------------------------
To unsubscribe, e-mail: cactus-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: cactus-user-help@jakarta.apache.org




Mime
View raw message