Ben Souther wrote:
Ah Ben, I don't know if you have kids or not. But y'know how a kid can kinda look at the floor and shuffle their feet when caught doing something stupid. Well, keep that in mind as you read what I figured out...I just dropped your JSPs in a box running win2k server and tomcat 5.0.24.
They run fine, same session. Once the strings are created, they stay created, no nulls.
I'm hitting from a linux box using Mozilla, but I also tried from MSIE on the machine that's hosting your JSPs.
All looks good.
For kicks, try doing this: In MSIE, click Internet Options -> Privacy (tab) -> Advanced (button) Then check the "Always Allow Session Cookies" checkbox. Let me know if that fixes it.
-good luck
Recall that I first installed the latest non-alpha code 5.0.24 but without the fix. I think that is key to what I was seeing. Also I am running this Zone Labs Integrity Desktop thingie on Win2K that our IT dep't thinks will keeps us dumb users safe. So, I deploy to the local TC5 instance and see some odd problems that I've tried to nail down. Seemed like session data was not being kept indicating a new session on each request -- kinda like cookies were blocked, hmmmm. Investigation of both the IE & Mozilla browsers did not indicate that session cookies should be blocked, in addition the Zone labs thing is set to allow session cookies but block persistent and 3rd party cookies. Should be fine you'd think. So for the rest of my struggles I ignored the cookies settings.
In part of the struggles I installed the 5.0.25-alpha version of TC5.
Then in the end I allowed all cookies to be passed, actually turned of all security junk in the browsers and the zone labs thing, deleted all cached pages & cookies, etc. Then it worked! I then gradually turned everything back up, and it still works, woohoo!
What I don't understand is why it (5.0.24) seemed to be failing locally but ran OK from the Linux box. BTW, the FC1 box is now at the 5.0.25-alpha version and seems to be just peachy.
Soooo, I think it is possible that my local deployment woes stemmed from some strange cookie that 5.0.24 inserted? Could that be true though I must admit that I read but don't wholly understand the implications of defect.
In any case I'm tickled that it's working and I'm off to solve some of my 'real' runtime woes in our product (which TC has already pointed out at least one).
Many, many thanks to you Ben and others on the mailing list for their patience and assistance. I truly do appreciate your interest and help with this problem.
Kind regards,
Mike
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
