This is a step toward upgrading WebRender, which will be upgraded and
patched in the `third_party` directory. This change vendors the current
private branch of WebRender that we use and adds a `patches` directory
which tracks the changes on top of the upstream WebRender commit
described by third_party/webrender/patches/head.
This function is already quite big and the changes necessary for
properly laying out independent formatting contexts next to floats will
make it even more unwieldy. Split the function in two and add a helper
for calculating the containing block for children and the margin. As
time goes on the independent formatting case is probably going to be
more like the replaced case anyway.
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Layout 2020: Optimize collapsible margin lookahead
Every time that we would lay out a block box that could collapse its top margin with its contents, we would do a lookahead to compute the resulting margin in order to place floats correctly.
The problem is that this lookahead could iterate several descendants, but then when laying these we would run the lookahead again.
This patch restricts the lookahead to boxes that either aren't collapsing their top margin with their parent, or that have 'clear' different than 'none' (since clearance prevents collapsing margins with the parent).
Since the lookahead stops iterating when it finds a box that doesn't collapse its top margin with its parent, or whose 'clear' isn't 'none', this should ensure that lookahead never handles the same box twice.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes do not require tests because there shouldn't be any change in behavior
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Every time that we would lay out a block box that could collapse its
top margin with its contents, we would do a lookahead to compute the
resulting margin in order to place floats correctly.
The problem was that this lookahead could iterate several descendants,
but then when laying these we would run the lookahead again.
This patch restricts the lookahead to boxes that either aren't collapsing
their top margin with their parent, or that have 'clear' different than
'none' (since clearance prevents collapsing margins with the parent).
Since the lookahead stops iterating when it finds a box that doesn't
collapse its top margin with its parent, or whose 'clear' isn't 'none',
this should ensure that lookahead never handles the same box twice.
Update mozangle
This should allow servo to take advantage of faster compilation times on Windows.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Check computed (min-)block-size when determining margin collapse
To collapse margins through, CSS2 requires "zero or auto computed height" and "zero computed min-height" (the latter should also also include auto, which was introduced in CSS3 as the new initial value).
Similarly, "auto computed height" is required to collapse the bottom margin with the bottom margin of the contents.
Therefore this patch stops collapsing when the used height is clamped to zero by max-height, but the computed value is not zero. Same for non-zero percentages that are indefinite or that resolve to 0px.
Note that 0% and calc(0% + 0px) are still considered to be zero (so they allow margin collapse), this may a bit inconsistent but matches Firefox (for the collapsing through case).
This also matches the heuristics in find_block_margin_collapsing_with_parent(). We could change the heuristics instead, but then they would have to track block sizes in order to be able to resolve percentages.
The change makes margin-collapse-through-percentage-height-block.html fail, that test is only passing on Blink, which did a different interpretation of the spec. To be fair, various places of CSS2 loosely consider that indefinite percentages compute to auto, and even Firefox does so when determining whether to collapse the top margin with the contents.
Since the spec isn't particularly clear and interoperability is lacking, I filed https://github.com/w3c/csswg-drafts/issues/8919.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#29907
- [X] There are tests for these changes (1 test fails, see above)
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Resolve percentages in find_block_margin_collapsing_with_parent
<!-- Please describe your changes on the following line: -->
Padding percentages weren't being resolved, this function would think
that margins wouldn't collapse even if the percentage resolved to 0px.
And margin percentages were always resolved to 0px.
So now the function takes a ContainingBlock parameter that is used to
resolve these percentages properly.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [ ] These changes fix #___ (GitHub issue number if applicable)
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because ___
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Padding percentages weren't being resolved, this function would think
that margins wouldn't collapse even if the percentage resolved to 0px.
And margin percentages were always resolved to 0px.
So now the function takes a ContainingBlock parameter that is used to
resolve these percentages properly.
To collapse margins through, CSS2 requires "zero or auto computed height"
and "zero computed min-height" (the latter should also also include auto,
which was introduced in CSS3 as the new initial value).
Similarly, "auto computed height" is required to collapse the bottom margin
with the bottom margin of the contents.
Therefore this patch stops collapsing when the used height is clamped
to zero by max-height, but the computed value is not zero. Same for
non-zero percentages that are indefinite or that resolve to 0px.
Note that 0% and calc(0% + 0px) are still considered to be zero (so they
allow margin collapse), this may a bit inconsistent but matches Firefox
(for the collapsing through case).
This also matches the heuristics in find_block_margin_collapsing_with_parent().
We could change the heuristics instead, but then they would have to track
block sizes in order to be able to resolve percentages.
The change makes margin-collapse-through-percentage-height-block.html fail,
that test is only passing on Blink, which did a different interpretation
of the spec. To be fair, various places of CSS2 loosely consider that
indefinite percentages compute to auto, and even Firefox does so when
determining whether to collapse the top margin with the contents.
Since the spec isn't particularly clear and interoperability is lacking,
I filed https://github.com/w3c/csswg-drafts/issues/8919.
Layout 2020: Properly calculate clearance
<!-- Please describe your changes on the following line: -->
calculate_clearance() was not taking into account that adding clearance prevents top margin from collapsing with earlier margins.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#29885 and #29919
- [X] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
CSS viewport removal
It was removed from the spec and it's disabled everywhere.
This also removes the meta viewport support (which was implemented on top), but that also had a single test and is disabled everywhere, so I'm not too concerned, it can be implemented again if / when needed.
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes fix#26349
- [x] There are tests for these changes
layout_2020: Do linebreak for atomic inline-level elements
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes fix#29591 (GitHub issue number if applicable)
<!-- Either: -->
- [x] There are tests for these changes OR
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
This also removes the meta viewport support (which was implemented on top), but that also had a single test and is disabled everywhere, so I'm not too concerned, it can be implemented again if / when needed.
This directory now contains third_party software that is vendored into
the Servo source tree. The idea is that it would eventually hold
webrender and other crates from mozilla-central as well with a standard
patch management approach for each.
Properly position floats when subsequent boxes collapse margins with containing block (2)
PR #29939 tried to address this but missed various cases.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#29944
- [X] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Make Layout 2013 style a superset of Layout 2020 style
This is the first step toward removing the compile-time difference in
style between Layout 2020 and Layout 2013. This change:
- Adds `attr()` CSS content support to legacy layout
- Adds parsing support for some unsupported features such as `display: flow-root` and `display: contents`.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Clearance was implemented as a Length, where zero meant no clearance.
However, having a clearance of 0px should be different than having
no clearance, since the former can still prevent margin collapse.
This patch keeps the existing behavior, so it won't be possible to get
a clearance of Some(Length::zero()), but it prepares the terrain for
a follow-up to fix calculate_clearance to return the proper thing.
Properly position floats when subsequent boxes collapse margins with containing block
Margins should be able to collapse through floats when collapsing with
parent blocks (the containing block). To properly place floats in this
situation, we need to look at these subsequent floats to find out how
much of the margin will collapse with the parent.
This initial implementation is very basic and the second step would be
to cache this in order to avoid having to constantly recalculate it.
Fixes#29915.
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes fix#29915.
- [x] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Margins should be able to collapse through floats when collapsing with
parent blocks (the containing block). To properly place floats in this
situation, we need to look at these subsequent floats to find out how
much of the margin will collapse with the parent.
This initial implementation is very basic and the second step would be
to cache this in order to avoid having to constantly recalculate it.
Fixes#29915.
Co-authored-by: Oriol Brufau <obrufau@igalia.com>
Remove the DOMToTexture feature
This relies on WebRender's frame output API, `set_output_image_handler`,
which has been removed from the latest upstream [1]. It's sad to remove
this feature, which was probably a lot of work to implement, but it
seems difficult to patch WebRender to restore this functionality.
Fixes#29936.
1. https://hg.mozilla.org/mozilla-central/rev/361521e3c52324809553c555fb066d50f023d9bf
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes fix#29936.
- [x] These changes do not require tests because they remove functionality.
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
This relies on WebRender's frame output API, `set_output_image_handler`,
which has been removed from the latest upstream [1]. It's sad to remove
this feature, which was probably a lot of work to implement, but it
seems difficult to patch WebRender to restore this functionality.
Fixes#29936.
1. https://hg.mozilla.org/mozilla-central/rev/361521e3c52324809553c555fb066d50f023d9bf
css `inset` should work on Layout 2020
<!-- Please describe your changes on the following line: -->
`top`, `bottom`, `left`, `right`, are already implemented in layout-2020, so adding the shorthand can be enabled.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#29705 (GitHub issue number if applicable)
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because ___
I'm don't know if I need tests for this, how do I determine that?
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Fix clearance on replaced elements
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#29928
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because ___
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Update uuid
This is required to use the new version of gleam and update ipc-channels
without duplicating the dependency in the future.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Increase BFC root height to contain floats
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#29823
- [X] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Previously, final float positions were calculated when their parents
were positioned. This prevented proper positioning of absolute children
of floats with static insets, because they accumulate offsets as they
are hoisted up the tree.
This change moves the final float positioning to
`PlacementState::place_fragment` for the float itself so that it happens
before any insets are updated for hoisted descendants. In addition to
simplifying the code, this makes it a bit more efficient. Finally,
floats are taken into account when updating static insets of hoisted
boxes.
Fixes#29826.
In #29897 I did the simple naive thing, but it wasn't entirely correct.
This patch tries to address the problems. In particular:
- Clearance should prevent margins from collapsing through if it
happens between them, as opposed to on the element that owns them.
- The margins of an element with clearance can still collapse through,
and collapse with other siblings as normal, but the resulting
margin can't collapse with the bottom margin of the parent.
Simplify layout of absolutes with static insets
Absolutes with static insets need to be laid out at their ancestor containing blocks, but their position is dependent on their parent's layout. The static layout position is passed up the tree during hoisting and ancestors each add their own offset to the position until it is relative to the containing block that contains the absolute.
This is currently done with a closure and a fairly tricky "tree rank" numbering system that needs to be threaded through the entire layout. This change replaces that system.
Every time a child is laid out we create a positioning context to hold any absolute children (this can be optimized away at a later time). At each of these moments, we call a method to aggregate offsets to the static insets of hoisted absolutes. This makes the logic easier to follow and will also allow implementing this behavior for inline-blocks, which was impossible with the old system.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] These changes do not require tests because it should not change behavior.
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Prevent margins from collapsing through when separated by clearance
This fixes#29884 and improves #29896.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#29884 and improve #29896
- [X] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Layout 2020: Correct rendering of floated root
Fix two issues around floating a root element:
1. In the StackingContext code handle the case where a root element is a Float fragment and not a Box fragment. This fixes a debug assertion failure in the css/CSS2/float/float-root.html test.
2. When initializing the SequentialLayoutState, use the containing block width as the maximum inline float placement position instead of infinity. This fixes the rendering of css/CSS2/float/float-root.html.
Note that css/CSS2/float/float-root.html was passing before, because both the test and reference were subject to the same bug. This fixes a couple other tests as well.
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [x] There are tests for these changes
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Fix two issues around floating a root element:
1. In the StackingContext code handle the case where a root element is a
Float fragment and not a Box fragment. This fixes a debug assertion
failure in the css/CSS2/float/float-root.html test.
2. When initializing the SequentialLayoutState, use the containing block
width as the maximum inline float placement position instead of
infinity. This fixes the rendering of css/CSS2/float/float-root.html.
Note that css/CSS2/float/float-root.html was passing before, because
both the test and reference were subject to the same bug. This fixes a
couple other tests as well.
Absolutes with static insets need to be laid out at their ancestor
containing blocks, but their position is dependent on their parent's
layout. The static layout position is passed up the tree during hoisting
and ancestors each add their own offset to the position until it is
relative to the containing block that contains the absolute.
This is currently done with a closure and a fairly tricky "tree rank"
numbering system that needs to be threaded through the entire layout.
This change replaces that system.
Every time a child is laid out we create a positioning context to hold
any absolute children (this can be optimized away at a later time). At
each of these moments, we call a method to aggregate offsets to the
static insets of hoisted absolutes. This makes the logic easier to
follow and will also allow implementing this behavior for inline-blocks,
which was impossible with the old system.
Typically, block-level contents are stacked vertically, so this was just
taking the maximum size among all contents. However, floats can be
stacked horizontally, so we need to sum their sizes.
