Bo, thanks for the comments. Let's move the discussion
to the developer list and see if they have some
comments.
Thanks,
Mike
>>From: "Bo Xu" <[EMAIL PROTECTED]>
>>Reply-To: [EMAIL PROTECTED]
>>To: [EMAIL PROTECTED]
>>Subject: Re: destroy/init sequence
>>Date: Sun, 06 May 2001 23:13:58 -0000
>>
>>>From: Michael Bergknoff <[EMAIL PROTECTED]>
>>>Does the container maintain session state across
>>>a destroy/init sequence? If so is it up to each
>>>servlet to serialize or delete any objects stored
>>>in the session? For debugging, is it possible to
>>>control when the destroy method is called?
>>>
>>>Thanks,
>>>Mike
>>>[...]
>>
>>
>>Hi :-) I am not sure:
>>- IMHO, I think HttpSession doesn't has much
relation
>> with one special MyServlet, session is "of"
>> webapps/container(not this webapp or that...),
i.e.,
>> session is cross-webapp, so session is loaded by
>> another classloader which is different from the
webapp
>> classloader which loads MyServlet. so even if the
old
>> webapp classloader is destroyed/GCed, the session
object
>> and its Class object and its classloader are still
there.
>>[...]
>>
>
>Hi :-) I find a mistake in my last email:-)
>
>with TC4.0b1(standalone, JDK1.3, winnt40):
>- session is "of" a special webapp, with the same
browser:
> % diferent webapp "has" different session
> % diferent MyServlet in the same webapp "has" same
session
>
>- when I updated MyServlet in webapp0, then MyServlet
will be
> reloaded, then the session of webapp0 will be a new
> one(before it is expired)
>
> is the above right? why does the container destroy
session
> of a webapp when this webapp's classloader is
updated?
>
>
>have a nice day! :-)
__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - buy the things you want at great prices
http://auctions.yahoo.com/
