Correctness improvements:
* UTF errors are handled safely per spec instead of dangerously truncating
strings.
* There are fewer converter implementations.
Performance improvements:
* The old code did exact buffer length math, which meant doing UTF math twice
on each input string (once for length calculation and another time for
conversion). Exact length math is more complicated when handling errors
properly, which the old code didn't do. The new code does UTF math on the
string content only once (when converting) but risks allocating more than
once. There are heuristics in place to lower the probability of
reallocation in cases where the double math avoidance isn't enough of a
saving to absorb an allocation and memcpy.
* Previously, in UTF-16 <-> UTF-8 conversions, an ASCII prefix was optimized
but a single non-ASCII code point pessimized the rest of the string. The
new code tries to get back on the fast ASCII path.
* UTF-16 to Latin1 conversion guarantees less about handling of out-of-range
input to eliminate an operation from the inner loop on x86/x86_64.
* When assigning to a pre-existing string, the new code tries to reuse the
old buffer instead of first releasing the old buffer and then allocating a
new one.
* When reallocating from the new code, the memcpy covers only the data that
is part of the logical length of the old string instead of memcpying the
whole capacity. (For old callers old excess memcpy behavior is preserved
due to bogus callers. See bug 1472113.)
* UTF-8 strings in XPConnect that are in the Latin1 range are passed to
SpiderMonkey as Latin1.
New features:
* Conversion between UTF-8 and Latin1 is added in order to enable faster
future interop between Rust code (or otherwise UTF-8-using code) and text
node and SpiderMonkey code that uses Latin1.
MozReview-Commit-ID: JaJuExfILM9
1. We will add more shape-like types in the future, so it's better to
use union to reduce the memory usage.
2. Those shape-like types are mutual exclusive, so we could use union to
wrap them.
Differential Revision: https://phabricator.services.mozilla.com/D2746
--HG--
extra : moz-landing-system : lando
It used to be this way because of lifetime issues (plus the shadow
datas were in RwLocks at some point IIRC). Now we guarantee that as long as the
element is away the cascade data is as well, so we don't need to thread it
around.
Differential Revision: https://phabricator.services.mozilla.com/D2768
--HG--
extra : moz-landing-system : lando
We only have this so that ::-moz-placeholder keeps serializing as
::-moz-placeholder, but I don't think anybody really cares.
Edge aliases ::-webkit-input-placeholder to ::-ms-input-placeholder at parse
time as well, as can be seen in:
```
let s = document.createElement('style');
s.innerHTML = `input::-webkit-input-placeholder { color: red };`;
document.body.appendChild(s);
document.body.innerHTML = s.sheet.cssRules[0].cssText;
```
And I think this is more consistent with what we do for CSS properties that are
aliases.
Differential Revision: https://phabricator.services.mozilla.com/D2595
MozReview-Commit-ID: 3ImDWamJhxh
This saves about 37 KiB of memory across the UA style sheets.
MozReview-Commit-ID: EoZnlmyWwxX
--HG--
extra : rebase_source : cd8ef0ba838618f9a4583b7d9896caa3a0602199
The '-moz-menulist-button' value currently behavies identically to the
'menulist-button' value. This is not implemented as an alias because later
patches in this patch series will change the behavior of our pre-existing
'menulist-button' value to more closely match what Chrome does.
The '-moz-menulist-button' value currently behavies identically to the
'menulist-button' value. This is not implemented as an alias because later
patches in this patch series will change the behavior of our pre-existing
'menulist-button' value to more closely match what Chrome does.
--HG--
extra : rebase_source : b66bf6427db5be2eb12f4e0aa36d22a4da46555a
Always assume allowed-for-all-content. There are a couple callers which weren't
doing that:
* A unit test -> removed.
* ComputeAnimationDistance: Used for testing (in transitions_per_property), and
for the animation inspector. The animation inspector shouldn't show
non-enabled properties. The transitions_per_property test already relies on
getComputedStyle stuff which only uses eForAllContent.
* GetCSSImageURLs: I added this API for the context menu page and such. It
doesn't rely on non-enabled-everywhere properties, it was only using
eInChrome because it was a ChromeOnly API, but it doesn't really need this.
Differential Revision: https://phabricator.services.mozilla.com/D2514
MozReview-Commit-ID: 4VOi5Su3Bos
And general Element logging. We now print all the attributes for comparison.
If this turns out to be too verbose we can change it to diff them or something.
Differential Revision: https://phabricator.services.mozilla.com/D2471
MozReview-Commit-ID: 1Gl9AumdnvZ
:root can't change without getting unbound from the tree so no fancy stuff other
than that needed.
This removes a lot of revalidation and attribute invalidation matching from the
Chrome, and looks like it should be a good idea in general.
Differential Revision: https://phabricator.services.mozilla.com/D2462
MozReview-Commit-ID: 9B0EO9teczi
We have a different order in nsCSSPropertyId for no good reason. The only
invariant there is that longhands come before shorthands, and shorthands before
aliases.
Luckily that's also an invariant that NonCustomPropertyId has, so we can reuse
them.
Differential Revision: https://phabricator.services.mozilla.com/D2463
MozReview-Commit-ID: 1hsQu6hmqiN
The patch at bug 1478391 comment 6 changed the way the math in Scrollbarbutton*
worked, which pretty surely caused this.
Restore the original order and math to be the same as before bug 1478391.
MozReview-Commit-ID: CK3iOqeX2NW
This builds on bug 1428676 and introduces StyleAppearance, which replaces the
NS_THEME_* constants.
Really sorry for the size of the patch.
There's a non-trivial change in the gtk theme, which I submitted separately as
bug 1478385.
Differential Revision: https://phabricator.services.mozilla.com/D2361
MozReview-Commit-ID: DiSmMWK7Krp
We use the same setup WR uses, which is checking-in the files.
But I think it's much better than keeping the two things in sync manually :)
When you add a new value, you need to add it to the rust source, then run the
command, but since it doesn't need to build the style crate (it uses the Rust
AST, doesn't build) there's no problem.
Differential Revision: https://phabricator.services.mozilla.com/D2354
MozReview-Commit-ID: DnXkNAzP54H
We use the same setup WR uses, which is checking-in the files.
But I think it's much better than keeping the two things in sync manually :)
When you add a new value, you need to add it to the rust source, then run the
command, but since it doesn't need to build the style crate (it uses the Rust
AST, doesn't build) there's no problem.
Differential Revision: https://phabricator.services.mozilla.com/D2354
MozReview-Commit-ID: DnXkNAzP54H
The values in the boolean context depend on its feature. For examples, in the
case of prefers-reduced-motion 'no-preference' means false and 'reduced' mean
true in the boolean context, whereas in the case of prefers-contrast
'no-preference' means false and other two values, 'high' and 'low' means true
in the boolean context. To support it we introduce a child struct of
nsCSSKTableEntry that has an additional field representing the value in the
boolean context and use it when we have no specified value in the media feature
(i.e. in the boolean context).
MozReview-Commit-ID: 79HiW8l5ous
I put Balance as the first item in StyleColumnFill because it is the default
value.
MozReview-Commit-ID: A4DVMLL6XBa
--HG--
extra : rebase_source : 8c8d07b724bcd49dd4f4057fcbf0ccd46ed591ff
NS_STYLE_COLUMN_COUNT_UNLIMITED is unused, so I remove it.
MozReview-Commit-ID: HLHLn9ZbkUY
--HG--
extra : rebase_source : 9bf00e1db00051c4454719e205633717fabde050
Summary:
This fixes a couple fuzz bugs and prevents special-casing <svg:use> even more in
bug 1431255.
Unfortunately not as many hacks went away as I'd have hoped, since we still need
to match document rules, see the linked SVGWG issues.
But blocks_ancestor_combinators goes away, which is nice since it's on a very
hot path.
Test Plan: WPT for style invalidation, covered by existing tests otherwise.
Reviewers: heycam
Tags: #secure-revision
Bug #: 1450250
Differential Revision: https://phabricator.services.mozilla.com/D2154
MozReview-Commit-ID: C4mthjoSNFh
Thought I had to update this as well, but nope. When basically any style changes
we already update transitions.
needs_transitions_update already handles the physical mapping changing by
checking whether any transition for the physical property remain there or not.
MozReview-Commit-ID: 6vKwal4yzRU
The setup is that AnimationValue only contains physical properties, and
we physicalize when building keyframes and transitions.
MozReview-Commit-ID: 9dI20N0LFrk
If we have a rotate axis whose length is extremely large, we will get an
infinite value, and its normalized vector is a zero vector, instead of an
unit vector, i.e. (x/inf, y/inf, z/inf) == (0, 0, 0).
The solution is: we scale the vector, so the length becomes a finite value,
and we could get a valid unit vector. Therefore, we use
a different normalization method, robust_normalize().
MozReview-Commit-ID: L8SteFe09aO
--HG--
extra : rebase_source : 4568c8bf906a9246e4ef13672a9ed541852b974a
In order to drop old euclid version, we still need to bump euclid for
plane-split and gfx/*. However, it needs more update and is not related to
this bug, so let's do that in other place. Here, we bump euclid to
0.18.1, and update style/values/generics/transform.rs for it.
MozReview-Commit-ID: JfNAxkR8wgs
--HG--
rename : third_party/rust/euclid/.cargo-checksum.json => third_party/rust/euclid-0.17.3/.cargo-checksum.json
rename : third_party/rust/euclid/Cargo.toml => third_party/rust/euclid-0.17.3/Cargo.toml
rename : third_party/rust/euclid/src/homogen.rs => third_party/rust/euclid-0.17.3/src/homogen.rs
rename : third_party/rust/euclid/src/length.rs => third_party/rust/euclid-0.17.3/src/length.rs
rename : third_party/rust/euclid/src/lib.rs => third_party/rust/euclid-0.17.3/src/lib.rs
rename : third_party/rust/euclid/src/macros.rs => third_party/rust/euclid-0.17.3/src/macros.rs
rename : third_party/rust/euclid/src/point.rs => third_party/rust/euclid-0.17.3/src/point.rs
rename : third_party/rust/euclid/src/rect.rs => third_party/rust/euclid-0.17.3/src/rect.rs
rename : third_party/rust/euclid/src/rotation.rs => third_party/rust/euclid-0.17.3/src/rotation.rs
rename : third_party/rust/euclid/src/scale.rs => third_party/rust/euclid-0.17.3/src/scale.rs
rename : third_party/rust/euclid/src/side_offsets.rs => third_party/rust/euclid-0.17.3/src/side_offsets.rs
rename : third_party/rust/euclid/src/size.rs => third_party/rust/euclid-0.17.3/src/size.rs
rename : third_party/rust/euclid/src/transform2d.rs => third_party/rust/euclid-0.17.3/src/transform2d.rs
rename : third_party/rust/euclid/src/transform3d.rs => third_party/rust/euclid-0.17.3/src/transform3d.rs
rename : third_party/rust/euclid/src/trig.rs => third_party/rust/euclid-0.17.3/src/trig.rs
rename : third_party/rust/euclid/src/vector.rs => third_party/rust/euclid-0.17.3/src/vector.rs
extra : rebase_source : 0594429c74c7391f80b4e291cd7fe9cbddd72b33
To be consistent, use the block indentation the rest of the Rust code uses.
Differential Revision: https://phabricator.services.mozilla.com/D1906
--HG--
extra : moz-landing-system : lando
... and cleanup unused keywords / getters using the scripts in
layout/style/tools
Differential Revision: https://phabricator.services.mozilla.com/D1901
MozReview-Commit-ID: BRUGcje7X0q
Most of it is automated by:
%s/eStyleContentType_/StyleContentType::/g
%s/nsStyleContentType/StyleContentType/g
But I removed some parentheses by hand.
Differential Revision: https://phabricator.services.mozilla.com/D1900
MozReview-Commit-ID: 3IcirjIYX5p
The 'all' shorthand has shipped a long time ago, so this pref is not needed
anymore.
MozReview-Commit-ID: GND8qSVAfCG
--HG--
extra : rebase_source : 10708e749911fa95554ed423a5782db61df67cd0
This changes the order of properties returned from gCS. The old order
doesn't make much sense, and other browsers don't agree on an identical
order either, so it should be trivial to change it. Also the spec isn't
super clear / useful in this case.
Several -moz-prefixed properties are excluded from the list due to their
being internal. I suspect they are never accessible anyway, so probably
nothing gets changed by this.
MozReview-Commit-ID: 9LfangjpJ3P
--HG--
extra : source : 879a7265c35f51c5954d8a44ccd374a606ecba0e
The EffectSet count does not exactly represent the count what we really need
for AnimationValueMap, but in most cases it matches. For example;
1) The element has two different keyframes animations
@keyframes anim1 {
to { opacity: 0; }
}
@keyframes anim2 {
to { transform: rotate(360deg); }
}
In this case the number matches.
2) The element has two animations but both keyframes have the same CSS property
@keyframes anim1 {
to { opacity: 0; }
}
@keyframes anim2 {
to { opacity: 0.1; }
}
In this case the number doesn't match, moreover it results more memory than we
ever needed, but this case is presumably less common.
3) The element has an animation having keyframes for two different CSS
properties.
@keyframes anim {
from { opacity: 0; transform: rotate(360deg); }
}
In this kind of cases, the number doesn't match. But even so, this patch
reduces the opportunities that the AnimationValueMap tries to allocate a new
memory (i.e. less opportunities on expanding the map).
Note that when the hash map is expanded, we do allocate a new RawTable with the
new size then replace the old one with the new one [1], so I believe this
change will reduce the crash rate to some extent.
[1] https://hg.mozilla.org/mozilla-central/file/15c95df467be/servo/components/hashglobe/src/hash_map.rs#l734
MozReview-Commit-ID: 6tcF9aqXh7a
--HG--
extra : rebase_source : 366989d3a2756f5a5711503a57f42f3b746d93a5