Serious IE bug: Workaround?

2 views
Skip to first unread message

Thomas Andrews

unread,
Jun 1, 2003, 7:56:26 PM6/1/03
to
Anybody have a workaround for this on IE?

http://thomaso.best.vwh.net/examples/ie/album/bug.html

In the version of IE I'm using, 6.0.2.2800.1106IS, the first
<h1> and <p> blocks are not displayed, but if you cover them with a
window and uncover, they are redrawn.

Oddly, without the "background: white", the page renders correctly.

Anybody seen this bug and have a workaround? In this case, of course, I
could remove the "background:" but in my original case, I cannot.

It doesn't help to give the floated element width and height styles.

It's an incredibly short example, with valid HTML 4.01 Strict code:

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
"http://www.w3.org/TR/html4/strict.dtd">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<title>Serious IE Bug</title>
</head>
<body style='background: white;'>
<div style="background: white; color:black;">
<h1>Serious IE Bug</h1>
<p>This is an example of an IE bug where, for some reason, a float
disrupts visibility of earlier blocks.</p>
<p style='float: left;'>Some text.</p>
<hr style='clear:both; visibility: hidden;'>
</div>
</body>
</html>


--
Thomas Andrews (tho...@best.com) http://thomaso.best.vwh.net/

Andrew Thompson

unread,
Jun 1, 2003, 8:32:41 PM6/1/03
to
"Thomas Andrews" <tho...@best.com> wrote in message
news:KowCa.9765$Nf.2...@sea-read.news.verio.net...

> Anybody have a workaround for this on IE?
>
> http://thomaso.best.vwh.net/examples/ie/album/bug.html
>
> In the version of IE I'm using, 6.0.2.2800.1106IS, the first
> <h1> and <p> blocks are not displayed, but if you cover them with a
> window and uncover, they are redrawn.

I'm sorry Thomas - I don't have any info about
your IE 6 bug. ..In 5.00 (Win2K) it appears fine.

What made me post, was that while both your pages
validate if I go to http://validator.w3.org/ and paste
the URL, if I click the link on the page the validator
responds..
"Sorry, this type of URI scheme (undefined) is not
supported by this service"

Can anyone explain to me what's happening there,
and how to correct it? I'd like to put the link
on test pages on my sites..

--
Andrew Thompson
http://physci.org/
http://lensescapes.com/

Thomas Andrews

unread,
Jun 1, 2003, 9:23:22 PM6/1/03
to
Andrew Thompson wrote:

The link uses the "referrer:" HTTP field. Maybe your browser isn't
sending referrers?

Iraq Reality Check: http://thomaso.best.vwh.net/Iraq.html

"We think the situation is far too serious, deadly and
dangerous to worry about what some Americans choose to call
their potatoes." - The French Ambassador

Andrew Thompson

unread,
Jun 1, 2003, 10:54:37 PM6/1/03
to
"Thomas Andrews" <tho...@best.com> wrote in message
news:eGxCa.9766$Nf.2...@sea-read.news.verio.net...

> Andrew Thompson wrote:
> > "Thomas Andrews" <tho...@best.com> wrote in message
> > news:KowCa.9765$Nf.2...@sea-read.news.verio.net...
>> ..In 5.00 (Win2K) it appears fine.
> >
> > What made me post, was that while both your pages
> > validate if I go to http://validator.w3.org/ and paste
> > the URL, if I click the link on the page the validator
> > responds..
> > "Sorry, this type of URI scheme (undefined) is not
> > supported by this service"
...

> The link uses the "referrer:" HTTP field. Maybe your browser isn't
> sending referrers?

I tested it using Mozilla 1.3 just then, same result..

The pages validate, just not if I click directly on
the link in the page, for either browser, I get..
" Sorry, this type of URI scheme (undefined) is not.."
in both.

(shrugs) Seems not to be an issue with any
_particular_ browser from this end.
Does it work for other people?

Thomas Andrews

unread,
Jun 2, 2003, 12:36:05 AM6/2/03
to

The other possibility is that you are srufing via a proxy which isn't
sending referrers, I guess.

Andrew Thompson

unread,
Jun 2, 2003, 1:17:35 AM6/2/03
to
"Thomas Andrews" <tho...@best.com> wrote in message
news:VuACa.9771$Nf.2...@sea-read.news.verio.net...
> Andrew Thompson wrote:
..

> >>>the URL, if I click the link on the page the validator
> >>>responds..
...

> The other possibility is that you are srufing via a proxy which isn't
> sending referrers, I guess.

Aaah. Thanks - I'll investigate it in another thread
(..and vacate this one, so hopefully you'll get some
answers from _your_ question here!!)

Karl Smith

unread,
Jun 2, 2003, 1:41:10 AM6/2/03
to
"Thomas Andrews" <tho...@best.com> wrote:

> > http://thomaso.best.vwh.net/examples/ie/album/bug.html
> >
> > In the version of IE I'm using, 6.0.2.2800.1106IS, the first
> > <h1> and <p> blocks are not displayed, but if you cover them with a
> > window and uncover, they are redrawn.

I can confirm the bug exists in IE6 on Win ME. I've seen it many times
before on my own files, but never on the web. Oh wait, I never use IE
on the web - that's why.


"Andrew Thompson" <Andr...@bigpond.com> wrote:

> ... http://validator.w3.org/ ...


> I'd like to put the link on test pages on my sites.

If you want the link to the validator for your own convenience, rather
than wanting to display the icon for whatever reason, consider using:

<link rel="validator" ...>

In the head, rather than an A in the body. Much less obtrusive.


--
And apparently I wrote to myself (though I don't recall doing so):

From: google-...@kjsmith.com
Reply-To: yourjob...@excite.com
To: google-...@kjsmith.com
Subject: If You Read Nothing Else - Read This
Date: Sat, 31 May 2003 17:50:40

Dear, Mr. or Mrs. google,

Andrew Thompson

unread,
Jun 2, 2003, 2:21:53 AM6/2/03
to
"Karl Smith" <google-...@kjsmith.com> wrote in message
news:3d18d2.03060...@posting.google.com...
..

> "Andrew Thompson" <Andr...@bigpond.com> wrote:
...
> If you want the link to the validator for your own convenience, rather
> than wanting to display the icon for whatever reason, consider using:
>
> <link rel="validator" ...>
>
> In the head, rather than an A in the body. Much less obtrusive.

I'll check further into that.

btw ..liked your 'best biewed using common sense'
link at http://users.tpg.com.au/karl6740/,
but I think I might try and wip something
up in JS (both test machines at least, will
have JS) that takes a parameter like
index.html?valid to (try to) write the
full links (including the document's own
URL) for single click validation..

Jim Dabell

unread,
Jun 2, 2003, 3:13:42 AM6/2/03
to
Karl Smith wrote:
> "Thomas Andrews" <tho...@best.com> wrote:
>
>> > http://thomaso.best.vwh.net/examples/ie/album/bug.html
>> >
>> > In the version of IE I'm using, 6.0.2.2800.1106IS, the first
>> > <h1> and <p> blocks are not displayed, but if you cover them with a
>> > window and uncover, they are redrawn.
>
> I can confirm the bug exists in IE6 on Win ME. I've seen it many times
> before on my own files, but never on the web. Oh wait, I never use IE
> on the web - that's why.

Even if you did use IE, you wouldn't see it - since it's in such widespread
usage, it would be web hara-kiri to let pages loose with that kind of bug
affecting them.

This particular bug can be fixed by setting { position: relative; } on the
parent element (or possibly an adjacent element too, IIRC). In this case,
setting it on the outer <div> element should work. Obscure as hell, but
that's the kind of level you have to stoop to if you want people to
actually be able to read your website :)


> "Andrew Thompson" <Andr...@bigpond.com> wrote:
>
>> ... http://validator.w3.org/ ...
>> I'd like to put the link on test pages on my sites.
>
> If you want the link to the validator for your own convenience, rather
> than wanting to display the icon for whatever reason, consider using:
>
> <link rel="validator" ...>
>
> In the head, rather than an A in the body. Much less obtrusive.

Nice idea. You could have a media="all" rule to { display: none; } it, and
have a user-stylesheet that overrides it too.

--
Jim Dabell

Thomas Andrews

unread,
Jun 2, 2003, 4:33:33 AM6/2/03
to
Jim Dabell wrote:

> Karl Smith wrote:
>
>>"Thomas Andrews" <tho...@best.com> wrote:
>>
>>
>>>> http://thomaso.best.vwh.net/examples/ie/album/bug.html
>>>>
>

> This particular bug can be fixed by setting { position: relative; } on the
> parent element (or possibly an adjacent element too, IIRC). In this case,
> setting it on the outer <div> element should work. Obscure as hell, but
> that's the kind of level you have to stoop to if you want people to
> actually be able to read your website :)
>

Thanks. That's a very bizarre workaround.

Thomas Andrews

unread,
Jun 2, 2003, 4:51:03 AM6/2/03
to
Jim Dabell wrote:

> This particular bug can be fixed by setting { position: relative; } on the
> parent element (or possibly an adjacent element too, IIRC). In this case,
> setting it on the outer <div> element should work. Obscure as hell, but
> that's the kind of level you have to stoop to if you want people to
> actually be able to read your website :)
>

Sadly, that workaround doesn't fix my original example.

Sigh.

I even tried adding:

* { position: relative; }

to my stylesheet. With that, my page renders correctly, except that IE
doesn't activate the vertical scrollbar for the page, so you can only
see the top section of the page. Obvious solution: Get a bigger monitor.

IE is one brain-damaged piece of software. I suspect the best thing to
do if you "want people to be able to read your website" is to not use
floats. :-)

Thomas Andrews

unread,
Jun 2, 2003, 5:03:01 AM6/2/03
to
Thomas Andrews wrote:

> Jim Dabell wrote:
>
>> This particular bug can be fixed by setting { position: relative; } on
>> the
>> parent element (or possibly an adjacent element too, IIRC). In this
>> case,
>> setting it on the outer <div> element should work. Obscure as hell, but
>> that's the kind of level you have to stoop to if you want people to
>> actually be able to read your website :)
>>
>
> Sadly, that workaround doesn't fix my original example.
>
> Sigh.
>
> I even tried adding:
>
> * { position: relative; }
>
> to my stylesheet. With that, my page renders correctly, except that IE
> doesn't activate the vertical scrollbar for the page, so you can only
> see the top section of the page.

Okay, I've finally got a solution which seems to work:

body * { position: relative; }

Un-freakin-believable, but it works. Apparently, making the body's
position relative causes the scrollbar to not appear.

Safalra

unread,
Jun 2, 2003, 5:30:21 AM6/2/03
to
Thomas Andrews <tho...@best.com> wrote in message news:<KowCa.9765$Nf.2...@sea-read.news.verio.net>...

> Anybody have a workaround for this on IE?
>
> http://thomaso.best.vwh.net/examples/ie/album/bug.html
>
> In the version of IE I'm using, 6.0.2.2800.1106IS, the first
> <h1> and <p> blocks are not displayed, but if you cover them with a
> window and uncover, they are redrawn.

The text also appears if you select it (you'll see it white on blue).

> Oddly, without the "background: white", the page renders correctly.
>
> Anybody seen this bug and have a workaround? In this case, of course, I
> could remove the "background:" but in my original case, I cannot.

The same bug also appears (sometimes - I'll get a demo page done if I
have time) in IE6 if you have a list where one or more elements at the
end are 'empty' (that is, <li></li> with nothing in between) - the
entire list is then invisible, subject to the conditions described
above.

--- Stephen Morley ---
http://www.safalra.com

Chris Morris

unread,
Jun 2, 2003, 5:33:50 AM6/2/03
to
use...@safalra.com (Safalra) writes:

> Thomas Andrews <tho...@best.com> wrote:
> > Anybody have a workaround for this on IE?
> >
> > http://thomaso.best.vwh.net/examples/ie/album/bug.html
> >
> > In the version of IE I'm using, 6.0.2.2800.1106IS, the first
> > <h1> and <p> blocks are not displayed, but if you cover them with a
> > window and uncover, they are redrawn.
>
> The text also appears if you select it (you'll see it white on blue).

Also, resizing the IE window (in this example it needs to be *very*
narrow) can make the text reappear - it seems very dependent on the
relative heights of various elements.

--
Chris

Karl Smith

unread,
Jun 3, 2003, 5:19:08 AM6/3/03
to
"Andrew Thompson" <Andr...@bigpond.com> wrote:

> btw ..liked your 'best viewed using common sense'
> link at http://users.tpg.com.au/karl6740/,

Icon taken from John Walker's excellent personal site:

http://www.fourmilab.ch/


> but I think I might try and wip something
> up in JS (both test machines at least, will
> have JS) that takes a parameter like
> index.html?valid to (try to) write the
> full links (including the document's own
> URL) for single click validation..

This might interest you:

http://www.squarefree.com/bookmarklets/validation.html

--

Jacqui or Pete

unread,
Jun 3, 2003, 7:13:07 AM6/3/03
to
In article <3d18d2.030603...@posting.google.com>, google-2003-
0...@kjsmith.com says...

> "Andrew Thompson" <Andr...@bigpond.com> wrote:
>
> > btw ..liked your 'best viewed using common sense'
> > link at http://users.tpg.com.au/karl6740/,
>
> Icon taken from John Walker's excellent personal site:
>
> http://www.fourmilab.ch/
>
...

> > URL) for single click validation..
>
...
> http://www.squarefree.com/bookmarklets/validation.html
>
>
and http://freshmeat.net/releases/108362/
--
Your site's most important visitors are completely blind,
totally deaf, and use browsers you've never seen.
Reply all
Reply to author
Forward
0 new messages