lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: lynx-dev Problem with META HTTP-EQUIV="refresh". Is it a bug?


From: David Woolley
Subject: Re: lynx-dev Problem with META HTTP-EQUIV="refresh". Is it a bug?
Date: Tue, 20 Mar 2001 22:56:08 +0000 (GMT)

> 
> ************************************
> I want to use lynx to get to the Bank of America web site:
> https://money.bankamerica.com/cgi-bin/SingleSignonServlet
> 
> The html document returned from the server has a form for input of
> user ID and a password. But it also has a refresh command

This is yet another new way of breaking a site with Javascript.
Javascript is the real problem here.

> in it: <META HTTP-EQUIV="refresh"
> CONTENT="0;URL='/help/un-supported_browser.html">

Using Refresh to achieve a redirect, in particular using it with a timeout
of zero is given a SHOULD NOT status in the HTML 4.01 standard.  Refresh is
not specified by any IETF or W3C specification (notwithstanding the
warning not to abuse it in the HTML specification).  It is a Netscapism.

Using a non-absolute URL violates Netscape's specification for this
feature.

Lynx takes the view that the user should be allowed to read things at
their own pace, so treats it as a pseudo-link.

> If I use IE, I type in the user ID and password, click the submit button
> and get to the next page. In lynx, clicking the submit button brings me
> to the highlighted link of the refresh URL only. I checked with the

This is a double fault condition; Lynx doesn't support Javascript but
also doesn't handle Refresh the way they expect.  If you access this
page on IE 5.01, they way I have it configured, it goes a page telling
you to upgrade to at least IE 4.01 (or various Netscape versions)!
That's because they haven't considered that security conscious people
might not permit JavaScript to run even on recent browsers.

They also haven't considered that even early GUI browsers might not 
act on the "redirect".  As a result, browsers that are able to support
the page may actually generate a "please upgrade" and browsers that cannot
support the site may display the form!

I haven't looked at the code in detail, but I suspect that Javascript
with a suitably powerful object model is used to abort the refresh
on browsers that support it, and the refresh is used to to try to
dump all other users into the "please upgrade" page.

I presume that the submit button doesn't work without Javascript.
One common trick with links and image links is to have href="#"
and do the real work with onclick.  Maybe the button isn't really
a button but an image link (although the page took an annoying
time to load before redirecting, it redirected as soon as it was
fully formed, so I didn't get a good look).

> What is wrong here? Is it a bug, or I am doing something wrong?
> I use the latest version dev19 compiled for cygwin.

You are making the mistake of assuming that commerce cares about 
accessibility or anything other than the latest versions of the big 2,
configured as out of the box.

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]