23092 - Attempting to log in causing the Javascript engine to lock & fail

Status: RESOLVED WORKSFORME

(Core :: JavaScript Engine, defect, P3)

Description

[Roger Davenport]

When attempting to log into http://mail.excite.com (a free e-mail service), the
javascript causes the browser to jump to 100% cpu usage for about 30sec, thus
locking any user input.  Also, the javascript never submits the form data as
programmed.

Comment 1

[Mike McCabe]

This sounds like an interesting problem, but I don't see anything that indicates
what the problem is.  I'm marking this bug INVALID for now; feel free to reopen
it if you can provide a *simple* reproducible case that demonstrates the
problem.

Please see the Bugzilla bug writing guidelines, or Rob's notes on reporting bugs
against JavaScript at http://www.mozilla.org/js/tests/library.html#PLEASE

Thanks!
Mike

Comment 2

[Roger Davenport]

I'm sorry.. here's simple.
go to http://mail.excite.com
type "dingodaddy" into the "Member Name"
type "schmoe" into the "Password"
press the "Sign in" button
see mozilla FREEZE

Comment 3

[Mike McCabe]

Reassigning to browser-general.

It'd be a great help towards actually fixing this if you could identify just
where the problem occurs.  I.E. at what point in the excite webpage does Mozilla
freeze?  Is it possible to construct a much simpler webpage that still causes
the same problem?  A good, simple testcase (not some commercial site) makes it
*much* easier for us to identify the problem and get the right person looking at
it.

Comment 4

[Roger Davenport]

I'm sorry.. I thought it was pretty clear that mozilla freezes when you hit the
"Sign in" button.  The JAVASCRIPT freezes it.  Goes into a bad loop.  NO
debugging messages about it, either.

The only visible javascript is thus (as you can see, they include a file as
well, which I am seem to be forbidden to dig into now):

<script language="javascript">

<!--
var JauthStarted =0;
function ResetJauth()
{
JauthStarted =0;
}
// -->

</script>
<script language="javascript">

<!--
function reportError(msg,url,line) {
var str = "" + " Jscript err: " + msg + " on line: "
+ line + " URL: " + url + " Browser: "
+ navigator.appName + " " + navigator.appVersion;
document.loginbox.jerror.value = str;
return true;
}
window.onerror = reportError;
// -->

</script>
<script language="javascript" src="/javascript/jauth_n23.js">

</script>

I don't have the mozilla toolset installed to dig any deeper into this.  I just
thought you'd like a bug report.. so here it is.  I wish you would spend the 10
seconds to crank up the site and try it yourself.. or live with it.

You're choice.

Comment 5

[leger]

Clearing INVALID resolution due to reopen.

Comment 6

[John Morrison]

Using win95 2000010416, I do not get 100% CPU lockup. I do not have go to the
correct URL. A couple of notes:
 1) the "misdirected" URL may be a consequence of bug #14801 (multiple ';' in
    URL are parse incorrectly)
 2) The form onsubmit calls JAuth which among other stuff, implements the
    MD5 algorithm -- not your typical bit of js, but nothing that should
    cause a lockup.
I'll look further into this at some point (but as I don't have Linux I may
not be able to pin this down).

By the way, [e-mail] ... what build are you running (M12?, a
nightly build?, a CVS pull and compile?)

Comment 7

[Mike McCabe]

This also WORKSFORME, on a recent Linux build.

Still noticing a problem?  When is your build?

Comment 8

[Roger Davenport]

OKay.. Mozilla M12, Linux/Alpha running RedHat 6.  As long as it's working on
intel, then I can be happy that eventually I or someone will fix it on alpha.

Thanks guys/gals.. you all do great work.  Thanks for taking a look at that for
me.

Comment 9

[Blake Ross]

reopening and marking Future...

Comment 10

[Mike McCabe]

Roger, this still happen?  I'd like to close the bug.

Comment 11

[Phil Schwartau]

Based on the comments above, resolving WORKSFORME -

Comment 12

[Phil Schwartau]

Oops - just noticed Blake's comment. Blake, is this still a bug for you?
If not, please mark this "Verified" if you get a chance - thanks 

Comment 13

[Phil Schwartau]

Updating QA contact -