That makes sense, but using the <@userreference> tag is not the same as checking for the cookie, because the <@userreference> tag will return a value if the cookie is not present.
I am absolutely certain I am using 065 of the client, and the servers. I have the client running on a single webserver, pointing to two app servers in a load group.
<@userreference> and <@arg _userreference> are the same and is not the issue.
I have 4 people trying to lose session, and it seems to be rocksolid now, since I have replaced <@userreferencecookies> with the actual Set Cookie clause.
Also, I have a test server, running the client and server on the same machine, which has the same reproducable problem, and the same workaround fixed it also. Also confirmed 065.
When I upgrade from previous versions, I uninstall, and do a clean install, and just move in my ini files manually. So it is clean.
Robert.
On Jan 20, 2004, at 9:01 AM, Robert Shubert wrote:
Robert,
You can not retrieve the Witango_UserReference cookie value with the
@VAR tag. This was explained to me by Phil at one point which boiled
down to the fact that the value was the name or key of the userreference
and was processed before other values were set. <@USERREFERENCE>
produces the result.
Please double check that you are using .065 both server and plugin/cgi as corrections to cookies were made in versions between .062 and .065
Also, double check that <@USERREFERENCE> and <@ARG _userreference> are
always the same, you might have an issue there. Because the processing
order was switched to Cookie > SEARCHARG > POSTARG, it is now better to
not pass _userreference unless you are sure that session cookies are not
supported or that you are switching domains, where it is required.
Robert
-----Original Message----- From: Robert Garcia [mailto:[EMAIL PROTECTED] Sent: Tuesday, January 20, 2004 8:56 AM To: [EMAIL PROTECTED] Subject: Re: Witango-Talk: Cookie Bug
One more thing.
In order to create my own "Cookie Check" method, I was doing some tests. If I set a simple cookie, like <@assign cookie$myTest "This is a
test.">, I can verify the cookie is set through my browser prefs, and then read it back with <@var cookie$myTest>.
However, If I verify that the Witango_UserReference cookie is set in my
browser, if I try to read it out with <@var cookie$Witango_UserReference>, I get nothing. I don't want to use <@userreference> because that will not necessarily verify if the cookie
is written.
Any ideas?
Robert.
On Jan 20, 2004, at 5:15 AM, Robert Garcia wrote:
I have been working through cookie issues and loss of state issues for
months, and I have been able to reproduce the problem. I am using 065
on windows by the way.
It seems that the <@userreferencecookie> tag is supposed to check the
instance of the userref either as a search arg, or in a cookie, and only write a cookie if none present.
However, sometimes, even with no userref in the search arg or cookie,
sometimes the cookie is not written ( this usually happens when a user
first hits the site). What makes it worse is that I use the <@userreferenceargument> in every link on the site, and since it gets
created on the first hit, and the cookie didn't get written, the cookie definitely doesn't get written in subsequent hits, because the
search arg userref is always there.
As a quick test I replaced the default header:
HTTP/1.1 <@HTTPSTATUSCODE> <@HTTPREASONPHRASE><@CRLF>Content-Type: text/html<@CRLF><@SETCOOKIES><@userreferencecookie><@CRLF>
With:
HTTP/1.1 <@HTTPSTATUSCODE> <@HTTPREASONPHRASE><@CRLF>Content-Type: text/html<@CRLF><@SETCOOKIES>Set-Cookie: Witango_UserReference=<@USERREFERENCE>;path=/<@CRLF><@CRLF>
This manually sets the cookie on every hit, and seems to solve all my
problems. Until I build a class to check first then write the cookie,
I will keep this, it doesn't seem to hurt performance to much.
This definitely seems to be a bug, and a pretty significant one. I am
super busy, but I will try to send this up to witango this weekend unless someone already has.
It would seem to me that it would be better to check if the cookie exists, and write it if it doesn't regardless if the search arg userref is there. I am thinking through how this may be affected if someone bookmarks a page with a search arg userref, and then uses it.
So I am going to work on a method, any thoughts would be great._______________________________________________________________________
--
Robert Garcia President - BigHead Technology VP Application Development - eventpix.com 5910 Clark Rd Suite G Paradise, Ca 95969 ph: 530.645.4040 x222 fax: 530.645.4040 [EMAIL PROTECTED] - [EMAIL PROTECTED] http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/
_ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
--
Robert Garcia President - BigHead Technology VP Application Development - eventpix.com 5910 Clark Rd Suite G Paradise, Ca 95969 ph: 530.645.4040 x222 fax: 530.645.4040 [EMAIL PROTECTED] - [EMAIL PROTECTED] http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/
_______________________________________________________________________ _
TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
_______________________________________________________________________ _
TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
--
Robert Garcia President - BigHead Technology VP Application Development - eventpix.com 5910 Clark Rd Suite G Paradise, Ca 95969 ph: 530.645.4040 x222 fax: 530.645.4040 [EMAIL PROTECTED] - [EMAIL PROTECTED] http://bighead.net/ - http://eventpix.com/ - http://theradmac.com/
________________________________________________________________________ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
