There are cases where stylesheets are used partially during document load as you are doing progressive display of a page and the stylesheet comes in as you are displaying. I have seen things *like* the H1's being the color in the stylesheet but the text not. This is bad - you shouldn't use the stylesheet until it's all in. However, a much easier way to replicate this bug occurs on my homepage: http://www.fas.harvard.edu/~dbaron/ . You may want to make a local copy of the page and the preferred stylesheet, but not the alternates to replicate this (because that's roughly the way it works from here). The existence of bug 2765 also helps to see this bug in this case. What happens is that the page is displayed without stylesheets, but if I click on a link before the stylesheets load, the link turns the active color I give in the preferred stylesheet, even though the stylesheet has no other effects. (clicking on the link doesn't work either, but that's another problem...) I think you need to somehow withhold the sheets from the progressive display mechanism until they are fully loaded.

A single style sheet is never used until it (and all of its @imports) are fully loaded. When a page links multiple sheets, we currently will use each sheet as soon as its loaded, the we reflow the whole doc again after the last sheet is in. Perhaps what you're seeing is the effect of different sheets coming in during the initial reflow. If so, are you suggesting that NO linked sheets apply until they ALL are in?

Yes. That is, you shouldn't use a stylesheet at all until you've reflowed the whole document with it. Otherwise things get messy, and possibly illegible (if the sheet were designed badly, which many are). Fixing bug 2765 will fix the alternate part, but not the multiple (persistent) stylesheets problem. (I'm not sure I've seen that, though.) Perhaps I've only seen this on pages with alternate or multiple SS. That makes sense. So I guess you should either reflow each time you get a stylesheet completed or you shouldn't use them until they're all in. If positions are held on reflow, then it won't be so bad to reflow more. I think troy (or someone) has a bug on that.

This bug will probably be fixed by Peter forthcoming stylesheet manager class. Related bugs: bug 2029, bug 2765, bug 3421, bug 4529. To find related bugs, search the description field for the [SSMC] marker.

See also bug 5318.

Using 6/14 Apprunner, style is applied as a whole. Verifying bug fixed.