On Tue, 2 Mar 2004, Robert O'Callahan wrote:
>
> Currently we don't place any bounds on parsed CSS values. Why not? I
> know the spec doesn't have any bounds, but of course there are
> "implementation specific limits"; we don't handle values > 2^31, for
> example, and other values get confused with preset constants like
> NS_UNCONSTRAINEDSIZE. Why don't we make things a bit more predictable,
> for example by constraining parsed CSS pixel values to be within plus or
> minus 10^7? (to pick a number out of thin air)

We should make sure that when we do this, values outside this range get
clamped to the range limit (or have some other well-defined mechanism
that isn't just dropping the entire rule).

Some browsers just overflow without checking and so e.g. with:

   text-indent: -10000em

...end up putting text in all kinds of random places.

-- 
Ian Hickson                                      )\._.,--....,'``.    fL
U+1047E                                         /,   _.. \   _\  ;`._ ,.
http://index.hixie.ch/                         `._.-(,_..'--(,_..'`-.;.'
_______________________________________________
mozilla-layout mailing list
[EMAIL PROTECTED]
http://mail.mozilla.org/listinfo/mozilla-layout

Reply via email to