T. J. Steed

Navigation

Skip navigation.

Search

Site navigation

Email conversation

FromT. J. Steed
ToMe
Subjectproblem with setExpression statement
Date9 December 2005 01:55
Here is a brief explanation of what I am trying.

First off, in html file A, I have an iframe which will load html B. Html
file b contains a table. I want the table's header row (and td's) to
remained fixed, like a freeze pain in Excel. So, I have been using the
expression(document.documentElement.scrollTop) statement in CSS as well as
(trying annything to work) setExpression("top",
"document.documentElement.scrollTop") within javascript. For some reason or
another it won't work. When I scroll in the iframe, the header row will not
remain put...almost like the expression isn't firing. Any ideas?

Instead of creating a table, I have been using just a div tag for
simplicity, however, it doesn't seem to work either. The weird thing is that
when I just the setExpression statement to adjust the height (according to
the iframe's size), it works like a charm.

Your help would be greatly appreciated. If I need to send you source files
or anything let me know. If I wasn't clear in my statement feel free to ask.

T.J.
FromMe
ToT. J. Steed
SubjectRe: problem with setExpression statement
Date9 December 2005 08:16
Attachmentdemo page showing a statically positioned row
T.J.

> First off, in html file A, I have an iframe which will load html B. Html
> file b contains a table. I want the table's header row (and td's) to
> remained fixed, like a freeze pain in Excel.
> So, I have been using the expression(document.documentElement.scrollTop)
> statement in CSS as well as (trying annything to work) setExpression("top",
> "document.documentElement.scrollTop") within javascript. For some reason or
> another it won't work.

Because it is non-standard garbage that only works in Internet Explorer. You
should use a more cross browser oriented approach, and then apply the IE
cludges to get it working in IE.

The good browsers (opera, safari, konqueror, iCab, firefox) should use
position: fixed; to hold the row in the desired position (relative to the
viewport). It is not possible to fix this relative to the table itself (even
if you use position:relative on the table itself, fixed position still
relates to the viewport, and Firefox refuses to make tables into positioned
element containers - grrr), so you will need to ensure you know the exact
position of the table.

This is easiest achieved by setting body, html, and table padding and margin
to 0px. You can then set the margin top on the table to a specific amount to
give you space for the fixed row. Use position:fixed on the row and position
it top-left.

Now, the IE bugs and rubbish. IE does not understand position:fixed so it
needs position:absolute - position:absolute does not remain static as the
page scrolls, so you need the ugly expressions. When an element inside a
table is positioned, IE always positions it relative to the table, even
though there is no positioning on the table itself. The position is then
offset by the margin on the body (that is a freaky bug).

At the moment, it will overlap the table in IE, so you need to add a margin
to the body (and remove it from the table), so that the table moves down,
but the positioned row gets incorrectly offset (to the correct position).
The expression used for top could fail for three reasons.

1. you used document.body.scrollTop - in strict mode you must use
document.documentElement.scrollTop

2. you did not use 'px' but you must in strict mode

3. expressions with scrolling offsets do not update (because of a bug)
unless you assign them to a variable - don't ask why, it just works ;)

I have attached an example - it is ugly, but if you want to get any more
complicated, you may have to resort to a completely DHTML-only solution.
Note that this is tailored to specific IE misbehaviours. I have no idea if
it will work in IE 7. It _should_ work using the same stuff as the other
browsers, but no doubt there will be some bug that makes it fail - probably
the table->container bug that it already suffers from here. You will have to
solve that problem when it occurs.


Mark 'Tarquin' Wilton-Jones - author of http://www.howtocreate.co.uk/
FromT. J. Steed
ToMe
SubjectRe: problem with setExpression statement
Date9 December 2005 14:19
First off, I greatly appreciate your help, especially the timely fashion in
which you responded. It means alot.

Now for the fun stuff. Unfortunately, because of the government contract
that we are on, the web application MUST run on IE 6, nothing else matters.
So, in some regards I guess that it is a bit easier since we don't have to
code for multi-browser support. However, as you know, it just happens to be
the worst browser (IE in general) to support...LOL. Oh well. You win some
and you lose some, I guess.

As for the setExpression statement, I am still having some problems. 1) For
some reason or another, and it may just be the security on our system, I
didn't get the attachments/examples that you sent. Considering my level of
confusion (trust me, it is huge right now), any type of example would be
great (if possible...I am not trying to force you into anything...I am truly
sorry). 2) I am having real trouble setting a variable and attaching 'px'
within the setExpression statement (again, this is my own incompetance). My
understanding is that the setExpression statement takes two parameters, both
of which are string. So, I have tried the following with nothing seeming to
work:

document.getElementById("text").style.setExpression("top", "(myvar =
(document.documentElement.scrollTop + 100))" + "px");

or

document.getElementById("text").style.setExpression("top", "(myvar =
(document.documentElement.scrollTop + 100)) + 'px'");

or

document.getElementById("text").style.setExpression("top", "(myvar =
(document.documentElement.scrollTop + 100)) + px");

or

document.getElementById("text").style.setExpression("top", "(myvar =
document.documentElement.scrollTop + 100)" + "px");

and most permutations in between...lol.

All of these produce javascript errors. 3 and finally) I am having some
problems with the getElementById statement. Here is my issue. If I give my
iframe an id (let's say "scrollwin"), I can't get the iframe's document's
scrollTop using the following code (fired using a button):

document.getElementById("scrollwin").document.documentElement.scrollTop

However, I can get the proper scrollTop if I use the following code:

scrollwin.document.documentElement.scrollTop

Needless to say, I am a bit confused as to when you need to use
getElementById and when not to. Again this is on IE 6. I am so confused on
the concept that I feel as though I am just hacking at solutions without
understanding the underlying principle. Any thoughts would help dearly.

Sincere thanks and apologies,
T.J.
FromMe
ToT. J. Steed
SubjectRe: problem with setExpression statement
Date9 December 2005 15:52
> Now for the fun stuff. Unfortunately, because of the government contract
that we are on, the web application MUST run on IE 6, nothing else matters.

While I know that your government behave beyond reason in that reguard, I
live forever in the hopes that they will one day see the light, and change
to something better ;) When that day comes, good code will not need to be
updated, and will make the transition easier. I always code by that
principle.

> I didn't get the attachments/examples that you sent.

Pasted as email content below - hope this gets through ...

> document.getElementById("text").style.setExpression("top", "(myvar =
(document.documentElement.scrollTop + 100)) + 'px'");

This looks like the most appropriate - I would have expected this to work.

> If I give my iframe an id (let's say "scrollwin"), I can't get the
> iframe's document's scrollTop using the following code (fired using a
> button):
>
> document.getElementById("scrollwin").document.documentElement.scrollTop

hrms - I think IE expects you to use something silly like Document instead
of document (upper case "D") - most other browsers use contentDocument. If
all else fails, give it a name instead of an id, and use:
window.frames["scrollwin"].window.document.documentElement.scrollTop

[source code of demo page]
FromT. J. Steed
ToMe
SubjectRe: problem with setExpression statement
Date19 December 2005 20:05
Quick question about the expression and setExpression statement. I am using
it to freeze/lock columns and rows in a table. In doing so, I have found
that it is lighter on the CPU to apply the expression to a tr's style
instead of each td when freezing a row. If I were going to do the same to a
column, I would think that it would be best to work with the <col> or
<colgroup> tags instead of each td, however, I can't seem to get it to work.
My expression statement, within the style attribute of the <colgroup> or
<col> tag (I have tried both) would look something like this:

left: expression((lockvar =
document.getElementById("tbl-container").scrollLeft - 2) + 'px')


Also, I was wondering if it would work on a TBODY since it would seem faster
to do it on a TBODY than each individual row? Any ideas.

T.J.

PS
The same expression, when applied to each td functionally works, however,
totally bogs down our resources. Right now, our users want the capacity to
hold up to 10000+ records (without paging).

PSS
Your suggestions on the iframe and freeze pane worked great...thanks.
FromMe
ToT. J. Steed
SubjectRe: problem with setExpression statement
Date20 December 2005 08:32
T.J.

> If I were going to do the same to a column, I would think that it would be
> best to work with the <col> or <colgroup> tags instead of each td, however,
> I can't seem to get it to work.

Unfortunately, the col element is slightly weird. It is not like a real
element at all. In good browsers, it is merely a placeholder sitting behind
the column, and it is not attached to the cell elements themselves. In IE
(and safari/konqueror), it cannot even do that correctly. I think your only
choice will be to do one cell at a time.

> Also, I was wondering if it would work on a TBODY since it would seem
> faster to do it on a TBODY than each individual row? Any ideas.

Assuming you want to use several sequential rows at the same time, yes, that
sounds sensible. Whether it will work or not is up to your experimentation,
but in theory it will.
This site was created by Mark "Tarquin" Wilton-Jones.
Don't click this link unless you want to be banned from our site.