With the newly supported fuzzy matching in our test-web runner, we can
now define the expected maximum color channel and pixel count errors per
failing test and set a baseline they should not exceed.
The figures I added to these tests all come from my macOS M4 machine.
Most discrepancies seem to come from color calculations being slightly
off.
While 788d5368a7 took care of better text
marker positioning, this improves graphical marker positioning instead.
By looking at how Firefox and Chrome render markers, it's clear that
there are three parts to positioning a graphical marker:
* The containing space that the marker resides in;
* The marker dimensions;
* The distance between the marker and the start of the list item.
The space that the marker can be contained in, is the area to the left
of the list item with a height of the marker's line-height. The marker
dimensions are relative to the marker's font's pixel size: most of them
are a square at 35% of the font size, but the disclosure markers are
sized at 50% instead. Finally, the marker distance is always gauged at
50% of the font size.
So for example, a list item with `list-style-type: disc` and `font-size:
20px`, has 10px between its start and the right side of the marker, and
the marker's dimensions are 7x7.
The percentages I've chosen closely resemble how Firefox lays out its
list item markers.
This commit is a three-parter that is hard to separate without breaking
marker rendering:
1. Any marker style that results in a string, except for a literal
string (e.g. `list-style-type: "@"`), should get the string ". "
appended. We forgot to do this for the alpha and roman types.
2. Instead of using the "pixel size rounded up" from a font and adding
an arbitrary 1 to that, we now use the exact pixel size for as long
as possible to improve our vertical positioning of markers.
3. Instead of always adding a "default marker width" to the marker
content width, we now only do this if we did not have text metrics
available (i.e. the marker style is not a text type). This greatly
improves horizontal positioning of text markers.
83b6bc4 went too far by forbidding SVGSVGElement from establishing a
stacking context. This element type does follow the behavior of CSS
boxes, unlike inner SVG elements like `<rect>`, `<circle>`, etc., which
are not supposed to be aware of concepts like stacking contexts,
overflow clipping, scroll offsets, etc.
This change allows us to delete overrides of `before_paint()` and
`after_paint()` in SVGPaintable and SVGSVGPaintable, because display
list recording code has been rearranged to take care of clipping and
scrolling before recursing into SVGSVGPaintable descendants.
`Screenshot/images/css-transform-box-ref.png` expectation is updated and
fixes a bug where a rectangle at the very bottom of the page was not
clipped correctly.
`Screenshot/images/svg-filters-lb-website-ref.png` has a more subtle
difference, but if you look closely, you’ll see it matches other
browsers more closely now.
FIXME: Rendering modifications to a list is sometimes not pixel-perfect
vs. reference (likely a bug). After this is fixed, screenshot
tests from this commit will likely fail + can be moved to
ref tests.
This test has flaked over the years, so let's add a screenshot test to
catch future regressions.
This copy of the test was taken from:
https://www.webstandards.org/files/acid2/test.html#top
Our CI infra does not support navigating to the "#top" anchor out of the
gate. So the intro section was removed from this copy so that we render
the happy face immediately.
We're able to efficiently draw repeated bitmaps through Skia, but for
backgrounds we only did so if the background was `repeat-x` _and_
`repeat-y`, and not if just one was set. This meant that for backgrounds
that were only repeating in one direction, we were taking the slow path.
Turns out that this slow path also produced graphical artifacts when
zooming in and out, so let's not do that :^)
Otherwise, the arrow painted next to the <details> element does not
update.
Using a screenshot test here because apparently the direction of the
arrow has no effect on the layout or paint trees.
Browsers such as Chrome and Firefox apply an arbitrary scale to the
current font size if `normal` is used for `line-height`. Firefox uses
1.2 while Chrome uses 1.15. Let's go with the latter for now, it's
relatively easy to change if we ever want to go back on that decision.
This also requires updating the expectations for a lot of layout tests.
The upside of this is that it's a bit easier to compare our layout
results to other browsers', especially Chrome.
Canvas text painting needs to support per-glyph font fallbacks, which
means we can't hand over responsibility for text shaping to Skia and
instead need to extract glyph paths from our own shaped GlyphRun.
This reverts commit a73cd88f0c.
Emitting SaveLayer for each paintable made rasterization a lot slower
on every website because now Skia has to allocate enormous amounts of
temporary surfaces. Let's revert it for now and figure how to implement
it with less aggressive SaveLayer usage.
When decoding data into bitmaps, we end up with different alpha types
(premultiplied vs. unpremultiplied color data). Unfortunately, Skia only
seems to handle premultiplied color data well when scaling bitmaps with
an alpha channel. This might be due to Skia historically only supporting
premultiplied color blending, with unpremultiplied support having been
added more recently.
When using Skia to blend bitmaps, we need the color data to be
premultiplied. ImmutableBitmap gains a new method to enforce the alpha
type to be used, which is now used by SharedResourceRequest and
CanvasRenderingContext2D to enforce the right alpha type.
Our LibWeb tests actually had a couple of screenshot tests that exposed
the graphical glitches caused by Skia; see the big smiley faces in the
CSS backgrounds tests for example. The failing tests are now updated to
accommodate the new behavior.
Chromium and Firefox both seem to apply the same behavior; e.g. they
actively decode PNGs (which are unpremultiplied in nature) to a
premultiplied bitmap.
Fixes#3691.
Instead of trying to manually determine which parts of a bitmap fall
within the box of the `<img>` element, just draw the whole bitmap and
let Skia clip the draw-area to the correct rectangle.
This fixes a bug where the entire bitmap was squashed into the rectangle
of the image box instead of being clipped.
With this change, image rendering is now correct enough to import some
of the WPT tests for object-fit and object-position. To get some good
coverage I have imported all tests for the `<img>` tag. I also wanted to
import a subset of the tests for the `<object>` tag, since those are
passing as well now. Unfortunately, they are flaky for unknown reasons.
This is the second attempt at this bugfix. The prior one was e055927ead
and broke image rendering whenever the page was scrolled. It has
subsequently been reverted in 16b14273d1. Hopefully this time it is not
horribly broken.
Instead of trying to manually determine which parts of a bitmap fall
within the box of the `<img>` element, just draw the whole bitmap and
let Skia clip the draw-area to the correct rectangle.
This fixes a bug where the entire bitmap was squashed into the rectangle
of the image box instead of being clipped.
With this change, image rendering is now correct enough to import some
of the WPT tests for object-fit and object-position. To get some good
coverage I have imported all tests for the `<img>` tag. I also wanted to
import a subset of the tests for the `<object>` tag, since those are
passing as well now. Unfortunately, they are flaky for unknown reasons.
Before this change, an element masked with 'mask-image: url(...)' would
show the mask, but 'mask: url(...)' would not. On e.g. dialogic.nl it
would show white boxes instead of the actual images in the top
navigation bar. We still do not support many of the other mask
properties, but with this change at least the masks show up in both
cases.
This has been left unimplemented since we switched to the Skia renderer.
Now `text-decoration-style: wavy` actually paints a wavy line. :^)
We had a text-decoration test, but it only checked `solid` lines, so
I've replaced it with a modified version of the old test page from
Serenity, without the blink option, and with some thickness parameters.
I did experiment with using a `SkPath1DPathEffect` to make it repeat the
pattern for us, but I couldn't make it look good at all.
By the time we're measuring the height of a BFC root, we've already
collapsed all relevant margins for the root and its descendants.
Given this, we should simply use 0 (relative to the BFC root) as the
lowest block axis coordinate (i.e Y value) for the margin edges.
This fixes a long-standing issue where BFC roots were sometimes not tall
enough to contain their children due to margins.
Using the raw value meant that 1em would be incorrectly treated as 1px,
for example.
I've updated our canvas-filters test to demonstrate this - without the
code change this would instead have an x-offset of 2px.
Previously most of the calculations for `object-fit` and
`object-position` were based on device pixels, meaning that images would
render differently based on zoom and DPI settings. Instead those
calculations now use css pixels and only the final draw-call is based
on device-pixels.
SVGs are rendered with subpixel precision. As such it can happen that
paths are rendered with less than 1px width or height and that they can
have a bounding box thinner than 1px. Due to an optimization such paths
were ignored when painting because their bounding box was incorrectly
calculated to be empty.
As a result horizontal or vertical lines inside SVGs were missing if:
* The SVG is displayed at viewbox size but the lines are defined with
less than 1px.
* The SVG contians 1px-thin lines, but is displayed at a size smaller
than viewbox size.
To prevent this, the bounding box of the path is now enlarged to contain
all pixels that are partially affected.
This improves the quality of our font rendering, especially when
animations are involved. Relevant changes:
* Skia fonts have their subpixel flag set, which means that individual
glyphs are rendered at subpixel offsets causing glyph runs as a
whole to look better.
* Fragment offsets are no longer rounded to whole device pixels, and
instead the floating point offset is kept. This allows us to pass
through the floating point baseline position all the way to the Skia
calls, which already expected that to be a float position.
The `scrollable-contains-table.html` ref test needed different table
headings since they would slightly inflate the column size in the test
file, but not the reference.
If not set, when copying pixels into the SkImage, skia assumes that the
color space is the same as the input, so no transformation is done. We
are currently setting the color space to sRGB, this is fine for now as
it allows us to start making some transformations, but down the road we
will want to set that to the actual output's display color space.
See previous the commit description for more details about the floating
points operations.
The hwb test cases in `css-color-functions` are now rendered identically
to what firefox does (I haven't checked the others tests, but they
aren't affected by this commit).
