Closed Bug 723370 Opened 12 years ago Closed 12 years ago

Bb Vista 8.0.6 Course Management System - Blank Screen with Firefox 10.0

Categories

(NSS :: Libraries, defect)

Product:
NSS
Component:
Libraries
Version:
3.13.1
Platform:
x86_64
Windows 7
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(firefox11- wontfix, firefox12- wontfix)

Status:
RESOLVED DUPLICATE of bug 702111
Tracking Flags:
Tracking Status
firefox11 - wontfix
firefox12 - wontfix

People

(Reporter: aengineerdu+computer, Assigned: briansmith)

References

Details

(Keywords: regression) 

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:10.0) Gecko/20100101 Firefox/10.0
Build ID: 20120129021758

Steps to reproduce:

Logged in to Drexel University's Course Management system Bb Vista 8.0.6.  The login worked fine, but I'm presented with a blank screen once I've entered the login information correctly.  Before the automatic upgrade to FF 10.0 this site worked fine.  
BBVista is a very widely used course management system.  This is a major problem.
I'm using FF 10.0 on a W7 64-bit machine with plenty of RAM.  I've never had a similar problems with previous versions of FF.  I checked with Drexel's support group.  They report that this problem has occurred for other users with FF 10 - it also happens with the latest version of Chrome.  I'm able to access the site with Opera and IE, both up to date.


Actual results:

Blank Screen after login.


Expected results:

I should have seen a page listing the courses for which I'm the professor.
We need either a attached testcase or public URL.
It quite possible that this is a bug in the page if it happens with chrome too
Please Contact Mr. Christopher Shields.  He can provide an account to test the system.


Chris Shields, MBA - cds28@drexel.edu
Director, Online Learning and Collaboration
Office of Information Resources and Technologies
Drexel University
215-895-1766
James: I sent an email to Christopher asking for a test account.
I just checked with Mr. Shields.  He said that he'd created the account (moztest) and enrolled it in a sample course.  He hasn't heard from you.

What's up?
Marcia: Do you received the Account details or should I request the Account ?
I'm a little bit busy at the moment and it could take a few days before I can look at this.
Hello James: I didn't an email back from Mr. Shields that I can see in my email. If you provide me with the URL and user name/password I will test this today.
Confirming - With the test account I see this using  Mozilla/5.0 (Windows NT 6.1; rv:10.0.2) Gecko/20100101 Firefox/10.0.2 as well as the latest nightly on Windows 7.

I tried Mozilla/5.0 (Windows NT 6.1; rv:9.0.1) Gecko/20100101 Firefox/9.0.1 and it works fine, so this is definitely a regression.

Will have to figure out what next steps are here.
Status: UNCONFIRMED → NEW
Ever confirmed: true
Keywords: regression, regressionwindow-wanted
Tracking for FF11/12 since this is a recent regression. Unless we find the root cause and we can back it out cleanly, this likely won't be resolved until FF12.

James - if this is a problem in both the latest Chrome and Firefox, it's quite possible that this is server side issue. Have you filed a bug with BB?
status-firefox11: --- → affected
status-firefox12: --- → affected
tracking-firefox11: --- → +
tracking-firefox12: --- → +
Since it worked with no problems with FF9 on the same version of VistaBB I figure that you folks changed something that is causing it to have problems now.  

I reported it to Mr. Shields of Drexel, who is our BBVista contact.  He stated that they (BBVista) are not interested in fixing this kind of problem since a new product is replacing it as of this coming fall. You're my only hope.

I've reverted to FF9 and am, of course, being nagged almost daily to upgrade - I wish I could.
Searching the exact regression range would help and you could try to safe the page and attach it.
Is that page using https ?
(In reply to Marcia Knous [:marcia] from comment #7)
> Confirming - With the test account I see this using  Mozilla/5.0 (Windows NT
> 6.1; rv:10.0.2) Gecko/20100101 Firefox/10.0.2 as well as the latest nightly
> on Windows 7.
> 
> I tried Mozilla/5.0 (Windows NT 6.1; rv:9.0.1) Gecko/20100101 Firefox/9.0.1
> and it works fine, so this is definitely a regression.
> 
> Will have to figure out what next steps are here.

Marcia/James - it would be very helpful to use mozregression with a BB Vista account (see [1]). Since this is working in FF9, the last known good nightly would be 2011-09-26.

[1] http://harthur.github.com/mozregression/
Adding qawanted for comment#11
status-firefox11: affected → wontfix
Keywords: qawanted
mozregression appears to be broken in the latest mozilla-build. I'm now trying to track down the regression window the good old fashioned way.
Last Good Version:
Firefox 10.0a1 Nightly 2011-10-07

First Bad Version:
Firefox 10.0a1 Nightly 2011-10-08
Keywords: qawanted, regressionwindow-wanted
[Triage Comment]
Glad we have a regression window. Looking at the pushes in that window there are a couple of inbound merges and bug 692520 looks like a potential related bug - Ehsan, can you look at this and confirm if we're on the right track here or if we should be looking to a different bug for the regression?
Can someone please convert the range in comment 14 to an hg range, using the revisions those two builds were built from?

Once we have that, somebody needs to bisect the range.  If bug 692520 is at fault here, then it's not obvious why.  The bisection should give us a better idea what's going on here.
Keywords: regressionwindow-wanted
(In reply to Ehsan Akhgari [:ehsan] from comment #16)
> Can someone please convert the range in comment 14 to an hg range, using the
> revisions those two builds were built from?

> Last Good Version:
> Firefox 10.0a1 Nightly 2011-10-07
Changeset: http://hg.mozilla.org/mozilla-central/rev/c3a50afc2243

> First Bad Version:
> Firefox 10.0a1 Nightly 2011-10-08
Changeset: http://hg.mozilla.org/mozilla-central/rev/6c780dcb4b99
Bug 689928 - White space in the left side of the tab bar [r=mbrubeck] 

Is that the info you were looking for Ehsan? If not, please direct me further and I'll be happy to help.
Yes, that's what I was looking for.  Still we need somebody who can reproduce this but to bisect this range.
(In reply to Ehsan Akhgari [:ehsan] from comment #18)
> Yes, that's what I was looking for.  Still we need somebody who can
> reproduce this but to bisect this range.

I was able to reproduce the issue -- how do I "bisect the range"? I've never done that before...
Assignee: nobody → ehsan
I can bisect if somebody from QA tells me exactly what I need to do in order to reproduce the bug.  :-)
Ehsan: When I was testing, after logging in I was not able to see anything on the screen using Firefox 10. If you compare with Firefox 9 you can see content on the page after logging in. So nothing really special to reproduce the bug other than a visual inspection.

(In reply to Ehsan Akhgari [:ehsan] from comment #20)
> I can bisect if somebody from QA tells me exactly what I need to do in order
> to reproduce the bug.  :-)
I've started to bisect the range...
The first bad revision is:
changeset:   78386:8f011395145e
user:        Brian Smith <bsmith@mozilla.com>
date:        Fri Oct 07 13:37:26 2011 -0700
summary:     Bug 669061: Upgrade to NSS 3.13 RC0, r=wtc

Brian, could you please see what change in the NSS upgrade has broken this?  I will forward the login information to the website to you in an email in a second.
Assignee: ehsan → bsmith
Blocks: 669061
Component: Untriaged → Libraries
Keywords: regressionwindow-wanted
Product: Firefox → NSS
QA Contact: untriaged → libraries
Version: 10 Branch → 3.13
Version: 3.13 → 3.13.1
It is very, very likely that this server is incompatible with the BEAST workaround; i.e. that this is a duplicate of bug 702111.

Note to QA: If a bug report is about a HTTPS site, and it is a regression in Firefox 10 or Firefox 11, and the user reports that Chrome has the same problem, then it is almost definitely bug 702111.

I verified this by running Firefox with NSS_SSL_CBC_RANDOM_IV=0 set in the environment. I was able to log in fine and I did not get the blank screen.

Please have your server software admins/developers read the comments in bug 702111; especially bug 702111 comment 35 (if your server is a Windows server) and *especially* bug 702111 comment 37.

I would not be surprised if your server software vendors have a fix ready for you to deploy now.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → DUPLICATE
You need to log in before you can comment on or make changes to this bug.