a separate `ClipDisplayItem`.
We push down clipping areas during absolute position calculation. This
makes display items into a flat list, improving cache locality. It
dramatically simplifies the code all around.
Because we need to push down clip rects even for absolutely-positioned
children of non-absolutely-positioned flows, this patch alters the
parallel traversal to compute absolute positions for
absolutely-positioned children at the same time it computes absolute
positions for other children. This doesn't seem to break anything either
in theory (since the overall order remains correct) or in practice. It
simplifies the parallel traversal code quite a bit.
See the relevant Gecko bug:
https://bugzilla.mozilla.org/show_bug.cgi?id=615734
splitting fragments.
I don't think it will be possible to avoid splitting fragments in the
presence of `vertical-align`, because one `ScannedTextFragment` could
potentially be split into arbitrary many fragments, each having its own
vertical position that can influence layout of other fragments.
This code also removes parts of `Range` that were no longer used.
stretching.
This preserves the usage of the Bloom filter throughout style recalc,
but the implementation is rewritten. Provides a 15% improvement on
Guardians of the Galaxy.
Note that using `servo --devtools http://example.org` doesn't work. In
that case either the port must be specified or the option moved to the
end. This is done by getopts and is the same for other such options,
e.g. `--profile`.
Previously, we implemented:
Au * Au -> Au
Au / Au -> Au
Au % Au -> Au
... which are inconsistent. It should be:
Au * Au -> SquaredAu
Au / Au -> i32
Au % Au -> i32
or:
Au * i32 -> Au
Au / i32 -> Au
Au % i32 -> Au
I picked the latter.
Also, the multiplicative identity does not make sense
when multiplication take two different types.
