The FontCache caches the result of font lookups. The cache key is a
simple object called FontSelector which consists of the font family
and font weight (both strings.)
This drastically reduces time spent in font lookup.
Some inline stylesheets use HTML comments like "<!--blah blah-->".
The HTML parser currently generates a comment child node of the <style>
element whenever this happens, and we don't want the comment itself to
be interpreted as part of the stylesheet.
This is a total hack to get around the auto-detection mechanism for
whether a block has inline or block children. We'll say that tables
never have inline children for now, and then anything that actually
turns out to be an inline child will just be ignored by layout.
Instead of computing whether a block's children are inline based on the
first child, make it an imperatively-set flag.
This gives us some flexibility to ignore things like text nodes inside
a <table>, for example. I'm still unsure what the "correct" way to deal
with those will be. We'll find out sooner or later. :^)
This class introduces LayoutTable, LayoutTableRow and LayoutTableCell.
These are produced by "display" values table, table-row and table-cell
respectively.
Note that there's no layout happening yet, I'm just adding the classes.
We now wait until the pixels are actually needed before fully decoding
images in <img> elements.
This needs some more work and is currently a bit memory-wasteful since
we'll hang on to the raw image data forever.
If a LayoutNode is split into line box fragments, we need to walk our
fragments and invalidate them. It was not enough to do this only for
LayoutBox nodes.
To streamline the layout tree and remove irrelevant data from classes
that don't need it, this patch adds two new LayoutNode subclasses.
LayoutNodeWithStyleAndBoxModelMetrics should be inherited by any layout
node that cares about box model metrics (margin, border, and padding.)
LayoutBox should be inherited by any layout node that can have a rect.
This makes LayoutText significantly smaller (from 140 to 40 bytes) and
clarifies a lot of things about the layout tree.
I'm also adding next_sibling() and previous_sibling() overloads to
LayoutBlock that return a LayoutBlock*. This is okay since blocks only
ever have block siblings.
Do also note that the semantics of is<T> slightly change in this patch:
is<T>(nullptr) now returns true, to facilitate allowing to<T>(nullptr).
This patch makes it possible to call Node::invalidate_style() and have
that node and all of its ancestors recompute their style.
We then figure out if the new style is visually different from the old
style, and if so do a paint invalidation with set_needs_display().
Note that the "are they visually different" code is very incomplete!
Use this to make hover effects a lot more efficient. They no longer
cause a full relayout+repaint, but only a style invalidation.
Style invalidations are still quite heavy though, and there's a lot of
room for improvement there. :^)
This is currently very aggressive. Whenever the Document's hovered node
changes, we invalidate all style and do a full relayout.
It does look cool though. So cool that I'm adding it to the default
stylesheet. :^)
Replaced elements will now properly create line breaks when they use up
the available horizontal space.
This fixes an issue with <img>'s lining up instead of breaking.
If you do a layout and it turns out that the page contents don't fit in
the viewport vertically, we add a vertical scrollbar. Since the
scrollbar takes up some horizontal space, this reduces the amount of
space available to the page. So we have to do a second layout pass. :^)
Fixes#650.
When style is invalidated (for example when an external stylesheet
finishes loading) we delete the whole layout tree and build a new one.
This is necessary since the new style information may result in a
different layout tree.
When layout is invalidated (window resized, image dimensions learned,
etc..) we keep the existing layout tree but run the layout algorithm
once again.
There's obviously lots of room for improvement here. :^)
The layout root is now kept alive via Document::m_layout_root.
This will allow us to do more layout-related things inside the inner
layer of LibHTML without reaching out to the HtmlView.
I'd like to keep HtmlView at a slightly higher level, to prevent it
from getting too complex.
This patch also fixes accidental disconnection of the layout tree from
the DOM after doing a layout tree rebuild. ~LayoutNode() now only
unsets the DOM node's layout_node() if it's itself.
This patch adds the CharacterData subclass of Node, which is now the
parent class of Text and a new Comment class.
A Comment node is one of these in HTML: <!--hello friends-->
Since these occur somewhat frequently on the web, we need to be able
to parse them.
This patch also adds a child rejection mechanism to the DOM tree.
Nodes can now override is_child_allowed(Node) and return false if they
don't want a particular Node to become a child of theirs. This is used
to prevent Document from taking on unwanted children.
The <br> element will produce a special LayoutBreak node in the layout
tree, which forces a break in the line layout whenever encountered.
This patch also makes LayoutBlock use the current line-height as the
minimum effective height for each line box. This ensures that having
multiple <br> elements in a row doesn't create 0-height line boxes.
We currently hard-code the line height to 140% of the font glyph height
and this patch doesn't fix the hard-coding, but at least moves it out
of LayoutText and into StyleProperties where it can be re-used until
the day we go and do a proper implementation of CSS line-height. :^)
This patch removes the hard-coded hack for "display: list-item" from
LayoutBlock and adds LayoutListItem and LayoutListItemMarker.
Elements with "display: list-item" now generate a LayoutListItem, which
may then also generate a LayoutListItemMarker if appropriate. :^)
For now this is simply a counter+hook exposed by ResourceLoader and
shown in the Browser status bar.
This is not very nuanced, and it would be nice to expose more info so
we could eventually do something like a progress bar.
This is definitely not the ideal ownership model here, but it's
something we'll have to iterate on as the engine grows.
At least this prevents us from leaking the entire world. :^)
We don't want to deal with document().frame() being null inside layout
tree code, so this makes sure we tear it down before the frame has a
chance to get nulled out.
