Closed Bug 26117 Opened 25 years ago Closed 24 years ago

HTTP Referer needs trailing / (Webbased email login failure)

Categories

(Core :: Networking, defect, P3)

defect

Tracking

()

VERIFIED WORKSFORME

People

(Reporter: mhearn, Assigned: jud)

References

()

Details

(Whiteboard: [NEED INFO][PDT-])

When I try and login (with correct password etc.) to 020 webmail I am simply
chucked back into the login screen when I click enter. It works with other
browsers and this is not related to my username/password which have worked fine
for months.
Reassigning to Eric.
Assignee: karnaze → pollmann
May also want to look at bug 25976 (can't log into excite webmail).
could be secure server issue.  Mozilla has the same problem with hotmail.com for
example
This is what we send to the server:

POST /cgi/checklogin.pl HTTP/1.0
Host: mail.020.co.uk
User-Agent: Mozilla/5.0 [en-US] (Linux; I)
Accept: */*
Accept-Language: en
Referer: http:///mail.020.co.uk
cookie: CSLAuthLR=http://mail.020.co.uk/020.html
Content-type: application/x-www-form-urlencoded
Content-Length: 112

referer=%2F020.html&config=system.cfg&CSL_domain=020.co.uk&submit2=ok&CSL_username=mozilla&CSL_password=password

Notice that we are sending the HTTP_REFERER:
Referer: http:///mail.020.co.uk

But Nav sends this HTTP_REFERER:
Referer: http:///mail.020.co.uk/

I modified our post to add the trailing / and things just magically worked.

Gagan, this seems like an HTTP issue - possibly related to bug 1582 - can you
take a look?  Thanks!
Assignee: pollmann → gagan
Severity: minor → normal
Component: Form Submission → Networking
OS: Windows 98 → All
Hardware: PC → All
Summary: Webbased email login failure → HTTP Referer needs trailing / (Webbased email login failure)
Keywords: beta1
Target Milestone: M14
Putting on PDT- radar for beta1.  Unless you can show broader impact.
Whiteboard: [PDT-]
warren says that this should probably be put in webshell along with http://
fixup... to jud. 
Assignee: gagan → valeski
Keywords: beta2
Target Milestone: M14 → M15
cc'ing andreas. We might be able to get away with arbitrarily appending a
trailing slash to HTTP urls in the webshell, but we can't do it for all urls.
FTP for example applies meaning to whether or not the *user* ended the url with
a slash.
Status: NEW → ASSIGNED
Someone please fill me in on the referer stuff. Why do we send 

Referer: http:///mail.020.co.uk

Why the three slashes? This would be a http url without a host and only a file
name mail.020.co.uk, which correctly has not trailing /.

If we would send http://mail.020.co.uk the parser would correctly add the
trailing / automatically. What is the correct syntax on the referer stuff?
Moving to M16.
Target Milestone: M15 → M16
this works now. I'm sure URL parsing changed such that we stick on the ending 
slash.
Status: ASSIGNED → RESOLVED
Closed: 24 years ago
Resolution: --- → FIXED
I don't think there were any changes to URL parsing. There must be another
explaination.
Keywords: nsbeta2
Reopening bug.

This bug still exists on:
 - MacOS9 2000-04-27-08-M16 Commercial Build
 - Linux6 2000-04-27-09-M16 Commercial Build
 - Win98 2000-04-27-09-M16 Commercial Build
Status: RESOLVED → REOPENED
Keywords: beta2
Resolution: FIXED → ---
Not sure if it's related, but when Mozilla follows a link to 
http://validator.w3.org/check/referer, it apparently does not send the referrer 
(or does not send it correctly). The W3C page opens up as if no referer had been 
sent. Last verified with 2000050820 on Windows 2000 Professional.
Is this blocking use of hotmail?
Whiteboard: [PDT-] → [NEED INFO][PDT-]
The http referer is working correctly now, per bugzilla#1582.

Marking VERIFED WORKSFORME on:
 - MacOS9 2000-05-18-09-M16 Commercial Build
 - Linux6 2000-05-18-09-M16 Commercial Build
 - Win98 2000-05-18-09-M16 Commercial Build
Status: REOPENED → RESOLVED
Closed: 24 years ago24 years ago
Resolution: --- → WORKSFORME
Marking Verified.
Status: RESOLVED → VERIFIED
Mass removing self from CC list.
Now I feel sumb because I have to add back. Sorry for the spam.
You need to log in before you can comment on or make changes to this bug.