See Firefox bug #87277.
Today (10 Dec, 2008), the CSS WG adopted an option on a css issue that seems to nullify the bug. See David Baron's comment, #22.
As I understand the issue, the difficulty lay in the anonymous box that encloses the table, and its caption. The diagram in section 17, tables, of the visual formatting model clearly illustrates the conundrum.
It seems the WG has clarified the model, making Firefox's rendering correct. Some browsers do manage to collapse table margins, but that may explain some of the issues with the way they handle captions, and why I seldom use captions, preferring
s or s.
If you'll look at comment numbers 16 and 18, you'll see that somebody finally listened to me. Or not.
An interesting set of
An interesting set of comments, do I remember reading this a while back? These bug tracks can make interesting reading and can be quite informative as to how things are supposed to work/behave.
I will take care though to never post a comment asking that things be hurried up and 'fixed' that comment met with a particularly frosty response, which I have memorised and am going to use here when people have unwarrented expectations
Uh, Mozilla doesn't like
Uh, Mozilla doesn't like anyone pointing out that their bugs often sit for a decade without getting fixed. Why I don't believe that automatically "open-source" = "all bugs are shallow" because that's BS. They're only shallow if someone decides to go fix it. Writing extensions must be so much more fun, because a lot of energy goes that direction.
Second, as I understand it there's no point in b*tching because they'll tell you to fix the bug yourself. Though if you do, you're a cowboy and not working within the community (and nobody likes you completely messing with the main trunk, which for these older bugs you usually have to do). Roll eyes. And you have to go from being a markup monkey to a C programmer? Gimme a decade.
Often the attitude is, "even though it doesn't work the same as all the other browsers in the universe, if the specs aren't crystal clear, then it's not a bug, it's a feature". The whole point of specs is so all vendors know what the correct behaviour should be of their product. Wherever something is "unspecified" you get everyone doing their own thing, and web developers losing their hair doing stupid hax to get all browsers acting the same... which was supposed to be fixed by having a standard.
Boris has said the unpositionable legend in FF is not a bug. The specs aren't clear enough for him (it doesn't clearly and explicitly state that legends should be able to act and be positioned like every other inline element).
Mozilla has also said the table positioning reference (the one where all other browsers can position something in reference to a table, not the one where table children can't position their children) is also not a bug, even though Gecko's the only one with the issue. I don't see the point in pointing at "it's undefined" in the specs when everyone else does it differently. All that does is ensure that Firefox will render a page completely differently than all the other browsers, and cause us to seek hacks.
If FF just had a simple * html- style hack, I'd happily ignore all its bugs, and just write around them. That's a tool I'd rather have built into the browser instead of letting decade-old bugs sit and wait for some C guru to come along an fix. What's more likely? That they'll get fixed? Or that we will continue writing web sites for all browsers and also for Mozilla? So the ugly solution is the better one. Let us have easy hacks. We've been doing it for IE for a decade now, right?
Poes you have a post
Poes you have a post compulsion, this is an old thread, it only surfaced due to some silly little spammer resurrecting it to post their trashy site links in which.