Thomas wrote: >http://neodude.net/host/cssd/fxselectbug.html
It looks like Donna's bug, but it seems different in practice since it definitely affects the mac version of FF1 as well. Plus, there's no scope for working round the problem - even shifting the excessive padding and margin on to generated content inside the block provides no relief. >I don't see a reasonable fix possible unless someone can "inspire" >Mozilla to put it at the top of their buglist. The more I look at the >technique, though, the more it seems like spacers revisited with a >modern twist... and I'm concerned about 32,000 of territory being part >of the DOM. Perhaps the envelope is being pushed a bit over the edge >(and into a 32,000px abyss)? 1. It's already fixed in Firefox 1.5. You may find the fact that it's broken in FF1.0 (and indeed as far back as Mozilla 1.6 - as for Moz 1.5, it won't even display the page content at all), a death knell for the fauxless equal height technique. On the other hand, I would argue that it has helped shake out a number of mistakes in the Gecko engine. 2. Really, Al. In what way, is a total separation of presentation and display and content code, like spacer gifs? And in what way is the faux columns technique less like spacers? Actually Al, let's have that discussion off list. As I've already said to you, write a critique of all or ant of the techniques and, unless there's some amazingly good reason not to, I'll add it to the article. 3. The article makes clear that the 32XXX figure shows the absolute limit you can push things to since beyond it, things will definitely go wrong. There's no reason to use such a large figure if you don't have to (which will be most of the time) Not that that changes anything really. Presumably you also take issue with absolutely positioning items and then setting their left value to -largenumber px. As for the 32XXXpx being part of the DOM - what do you mean? If things are implemented properly in the browser, the overflow: hidden on the containing box should make the browser behave as if there was literally nothing beyond its bottom edge. There *are* a couple of thorny problems that have arisen, one a printing problem in IE and the other relating to the use of anchors within an equal height columns layout - though as I am incredibly flu-ridden and don't have access to a full range of browsers to even confirm the problems let alone distill them to the point where I can pass them on the good folks of css-d to help out. Finally, just to stress this again, the "One True Layout" name is a gag. And if it refers to anything at all it is none of the techniques I illustrate, but rather the ability to stick to the task of designing a pure and semantically meaningful document structure. Style that resulting structure any way you like. Throughout the article I make it clear that in many instances pre-existing techniques will do just as well if not better than the ones I use there and even mention that javascript (unobtrusive of course) could be used instead if that is your preference. In the case of faux columns v fauxless columns... If you have a fixed-width pixel-based layout and are already using background images, using faux columns is probably a better bet since you can just combine them into one graphic. For percentage and em-based layouts (and liquid center, fixed sides) it probably holds true too, though that depends on how many columns you need and how many wrapping elements you're prepared to swallow. Let's face it, it's highly unlikely that we're talking about more than three which means you only need one additional wrapper. You could probably even claim it had some genuine semantic purpose ;) So to repeat the big take home message - there is no excuse to kludge your document structure. Lay your content out logically and accessibly. Otherwise I'm coming to get you. Alright? ______________________________________________________________________ css-discuss [EMAIL PROTECTED] http://www.css-discuss.org/mailman/listinfo/css-d List wiki/FAQ -- http://css-discuss.incutio.com/ Supported by evolt.org -- http://www.evolt.org/help_support_evolt/
