Pages w/JavaScript, InfoRefiner can delete all content (?)
Status: Beta
Brought to you by:
billthies
Menu
▾
▴
#15 Pages w/JavaScript, InfoRefiner can delete all content (?)
Opened by thies@mit.edu on 2002-03-06
This page has lots of images at the top, but there's
some content near the bottom that gets stripped out by
the filterer. The refined page seems to have all of the
image tags, but ends up displaying completely empty to
the client (no TEK disclaimer, either.)
http://www.coolplants.com/Tiki.htm
Is sending you the URL enough to reproduce these,
Tazeen? Can you just try running this page through the
filtering?
------- Additional Comments From thies@mit.edu
03/06/02 13:46 -------
Here's another page with content that ends up empty,
no frames:
http://www.ftd.com/pages/FTD110-1792.html
Might we be stripping away the content with the
Javascript removal? I think that might be it. The
refined page doesn't have empty source code, but is
displayed empty in the browser.
I'll stop reporting empty pages that have Javascript but
no frames. Let me know if you have a shortage of
examples.
------- Additional Comments From thies@mit.edu
03/06/02 14:07 -------
Sorry, just one more example. Since it's lucent's home
page, seemed more important:
www.ins.com
Again, no frames, but javascript, and empty in TEK
browser.
------- Additional Comments From tmahtab@mit.edu
03/14/02 18:26 -------
For http://www.ftd.com/pages/FTD110-1792.html, there
is a mistake in the
refining.
<script>
document.write("<link rel=stylesheet type='text/css'
href='/home/ftd.css?b=" +
browser_info + "'>");
</script>
</head>
<body bgcolor=#FFFFFF link=#0066CC alink=#0066CC
vlink=#0066CC >
is refined to
<link type =" text/css" href =" /home/ftd.css?b= +
browser_info + " rel ="
stylesheet">
</head>
<body _implied_ =" true">
");
However, when I open the refined file in Netscape, it
displays okay, with most of the content. When I look at
it through the proxy, it only displays a couple
of lines, and the page source is incomplete i.e. only a
fraction of the source in the actual file. This seems to
be another bug in the proxy.
In the future, when reporting on pages, can you first do
any/all of the following checks? It'd be really helpful in
distinguishing between server bugs
and proxy bugs.
1. The refined file in the cached dir opens in Netscape
and has content
2. The refined file opens in the proxy (the URL shows up
in the address bar)
and has content
3. Whether content of 1. and 2. are the same
4. If 2. doesn't work, whether clicking on the URL leads
to an intermediate
page (note URL in address bar) or whether it just loops,
trying to connect
5. If there is a Javascript error, type "javascript:" into
the address bar to
see what the error is
Bill, if you still have those other two pages around, could
you do this? I'm having a hard time figuring out whether
the errors I'm finding are the same as those you're
reporting.
Logged In: YES
user_id=574184
Needs to be tested with loband on server side.