Beware of Margins or Padding when Using the min-width Hack for IE
While we all watch as Internet Explorer 6 dies a slow, painful death, many unfortunate designers and developers continue to find themselves dealing with IE6’s lack of support for simple things like minimum and maximum widths. Fortunately, there are solutions to this problem, primarily in the form of CSS expressions such as this:
/* set the minimum width for IE 6 */
#target_element {
width: expression((document.body.clientWidth < 335)? "333px" : "auto"); /* min-width for IE6 */
min-width: 333px; /* min-width for all standards-compliant browsers */
}
Although ugly, invalid, and resource-intensive, this technique works well at setting min-width
properties for IE6. But beware! There is a catch. Chris Herdt recently wrote in with an important discovery:
…inclusion of any padding or margin on the element that has the fix applied will cause IE6 to crash…
After checking out Chris’ online demonstration of the behavior, I returned to the lab and began running a few tests. In addition to verifying and understanding Chris’ demonstration, I also wanted to know the following:
- Does this issue affect older (pre-6) versions of Internet Explorer?
- Is this issue seen in similar hacks such as max-width or max-height?
- What are some workarounds?
First of all, this issue seems only confined to Internet Explorer version 6. Although I didn’t test on IE4 or IE3 (why bother?), IE5 did not crash when put to the test. So, for all practical purposes, IE6 seems to be the only browser affected by this issue.
As for related CSS expressions, such as max-width and max-height, the addition of margins and/or padding does not seem to have such an effect. Only when using the CSS min-width
expression (i.e., hack) on IE6 will the addition of margins and/or padding crash the browser.
To explore a couple possible workarounds for this bug, let’s first review the mechanism by which the min-width
expression operates. Here it is again (so you don’t have to scroll up):
/* set the minimum width for IE 6 */
#target_element {
width: expression((document.body.clientWidth < 335)? "333px" : "auto"); /* min-width for IE6 */
min-width: 333px; /* min-width for all standards-compliant browsers */
}
The element for which we are applying the min-width
property is typically a div
. Here, we see that the ID of such an element is #target_element
, surprisingly enough. So we have this division, and we want it to remain at least 332 pixels wide, regardless of how the browser is resized1. To accomplish this in Internet Explorer, we must use the JavaScript expression seen in the example; however, to accomplish the same thing in all modern browsers (e.g., Safari, Opera, Firefox, IE7, et al), we need only specify a value for the min-width
property (i.e., the second declaration in our example). On a side note, this sort of browser-wrangling is exactly one of the reasons why people are so excited about the inevitable death of Internet Explorer 6.
So this is all fine and well and good, right? Well yes, it certainly works, but only until you try adding a bit of horizontal margin or padding to the target element, as seen in this example:
/* set the minimum width for IE 6 */
#target_element {
width: expression((document.body.clientWidth < 335)? "333px" : "auto"); /* min-width for IE6 */
min-width: 333px; /* min-width for all standards-compliant browsers */
margin-right: 1px; /* say goodnight, IE6! */
}
That little margin-right
declaration will crash IE6 in a bad way. In fact, any of the following horizontally effective declarations will crash IE6, regardless of their values:
margin-right: 1px;
margin-left: 1px;
margin: 0 1px 0 1px;
padding-right: 1px;
padding-left: 1px;
padding: 0 1px 0 1px;
This brings us to an easy workaround for this issue: don’t add horizontal (left and/or right) padding or margins to the target element of a min-width
expression. You can add all the vertical (top and/or bottom) padding and margins that you need, but stay away from teh horizontal stuff!
If you do need to apply space between the target element and its parent, consider using Chris’ workaround:
…I applied padding to the body instead of the content div and then just changed the expression to document.body.clientWidth – [padding values] < [desired minimum width +1].
Simply apply the padding (or margins) to the parent of the target element and subtract the difference from the CSS expression:
/* first apply the padding to the parent element */
#parent_element {
padding-right: 33px;
padding-left: 33px;
/* sum equals 66px */
}
/* then set the minimum width for IE6 accordingly */
#target_element {
width: expression((document.body.clientWidth - 66 < 335)? "333px" : "auto"); /* min-width for IE6 */
min-width: 333px; /* min-width for all standards-compliant browsers */
}
Of course, there are probably other workarounds to consider as well. Perhaps there is a way to configure the layout that doesn’t even require the min-width
property? I don’t know, but I do know one thing: if you are still required to design for Internet Explorer 6, you have my deepest sympathy.
Special thanks to Chris Herdt for discovering this issue and bringing it to my attention. With several articles now on the topic of using the CSS min-width
expression, it is important to share any pertinent information that may affect its implementation. Hopefully, this article will prove helpful to those encountering issues with the technique.
Footnote
1 When using the min-width
expression, always increase the value of your desired minimum width by (at least) 1 pixel. This is necessary to prevent Internet Explorer crashing, this time for a completely different reason. For more information, check out this comment.
15 responses to “Beware of Margins or Padding when Using the min-width Hack for IE”
@Cooltad: that looks like a good way of dealing with the maximum width scenario, but you would need to either decrease the first value of
900
by at least a pixel, or else increase the second instance of900
by at least a pixel in order for it to function properly. Likewise if you were to modify the script to account for the minimum width scenario, the first and second width values would need to differ in the same direction by at least one pixel for it to work. In any case, it’s another great tool to have available for dealing with good ‘ol IE6! ;)Wow, I made that script and totally forgot about the 1 pixel difference; thanks.
@Cooltad: yes, well, that information may have been less for you and others who are savvy and more for the other visitors who use this site as a technical reference for code examples, methods, etc. Anything I can do to alleviate confusion and unnecessary questions is a good thing! :)