Most of the change is moving sets around to be static functions on
LonghandIdSet. I think I like that pattern, but I can also make the new set a
global static and add mako code to be `pub` or something.
Though I think the LonghandIdSet::foo().contains(..) pattern is nice to read :)
Differential Revision: https://phabricator.services.mozilla.com/D10653
--HG--
extra : moz-landing-system : lando
When you're in a ShadowRoot and can share style with a sibling, the sharing code
is smart enough to skip document author rules.
But then it could get confused if you also include document rules, since
revalidation selectors are matched against these.
This is not a correctness issue, because we're matching more than what we need,
and avoid sharing if we failed.
Also fix the detection for user rules in any_applicable_rule_data.
Differential Revision: https://phabricator.services.mozilla.com/D10117
--HG--
extra : moz-landing-system : lando
It doesn't make much sense to return const UniquePtr<Foo>& for something that
can't be null, it's just confusing.
Also make more stuff actually const.
Differential Revision: https://phabricator.services.mozilla.com/D10647
--HG--
extra : moz-landing-system : lando
For now, we keep supporting the prefixed version, since there are examples/instructions
on the Web that don't include an unprefixed value.
Differential Revision: https://phabricator.services.mozilla.com/D10451
--HG--
extra : moz-landing-system : lando
A minor update to drop the redundant "mozilla" namespace prefix in
`cbindgen_types` array.
Depends on D10305
Differential Revision: https://phabricator.services.mozilla.com/D10325
--HG--
extra : moz-landing-system : lando
This needs to update the "fill-rule" and "clip-rule" to use
predefined_type to avoid some compilation errors.
Differential Revision: https://phabricator.services.mozilla.com/D10142
--HG--
extra : moz-landing-system : lando
We will blacklist this type and add a module raw line to map the gecko
type to its rust type (as an alias).
Differential Revision: https://phabricator.services.mozilla.com/D10303
--HG--
extra : moz-landing-system : lando
Before this patch we were only optimizing the case of a single selector, which
is fine, but not enough to catch ones like .foo .bar or so.
This patch allows us to optimize classes and tags in the rightmost compound,
while keeping the current optimization for #id selectors.
Need to profile this, but code-wise should be ready for review.
Differential Revision: https://phabricator.services.mozilla.com/D9351
--HG--
extra : moz-landing-system : lando
Also remove specified-value-only keywords, since those are handled
only in Rust code and C++ doesn't need to know about them.
Differential Revision: https://phabricator.services.mozilla.com/D9634
--HG--
extra : moz-landing-system : lando
The selectors that crash seem just corrupted data structures, none of the
selectors from crash dumps make sense, and the ones for which I could trace the
source found no issue.
1. Add a new preference, layout.css.step-position-jump.enabled, for
step(_, jump-*) timing functions.
2. We still keep JumpEnd and End tags, even though there is no difference
between them. Therefore, we could disable the preference if needed.
3. Update the calculation of StepTiming to match the algorithm in the spec.
4. For servo, we implement the correct step function algorithm except
for the handling of before_flag. This could be fixed later.
Depends on D9313
Differential Revision: https://phabricator.services.mozilla.com/D9314
--HG--
extra : moz-landing-system : lando
First, we generate StyleComputedTimingFunction by cbindgen from Rust, and use
it in nsTimingFunction, so we could copy it directly without handling
the different memory layout. However, we have to rewrite the
nsTimingFunction and mozilla::ComputedTimingFunction for this.
Second, the rust-bindgen seems cannot generate the correct generic members
from complex C++ templates, especially for the nested template struct,
(https://github.com/rust-lang-nursery/rust-bindgen/issues/1429)
So we have to hide StyleTimingFunction to avoid the compilation errors.
Depends on D9312
Differential Revision: https://phabricator.services.mozilla.com/D9313
--HG--
extra : moz-landing-system : lando
We make sure the step number is always positive, so using
computed::Integer is safe and can derive ToComputedValue.
Depends on D9311
Differential Revision: https://phabricator.services.mozilla.com/D9845
--HG--
extra : moz-landing-system : lando
TimingFunction is defined in a separate spec (i.e. css-easing), instead
of transform, so we move it into a different file.
Depends on D9310
Differential Revision: https://phabricator.services.mozilla.com/D9311
--HG--
extra : moz-landing-system : lando
frames() timing function was removed from the spec, so we drop it.
Besides, some devtool tests are removed because they use frame(). I will
add them back by using new step function later.
Differential Revision: https://phabricator.services.mozilla.com/D9309
--HG--
extra : moz-landing-system : lando
See the extended commit message for the following spec change:
6b36d41ebc
Basically, by failing to take the absolute value, for certain content we can end
up doing division by zero which will mean that the test included in this patch
will cause an assertion to fail in debug builds and return
"matrix(NaN, NaN....)" in release builds.
Differential Revision: https://phabricator.services.mozilla.com/D9618
--HG--
extra : moz-landing-system : lando
This change removes directives NS_STYLE_ANIMATION_PLAY_STATE_* and replaces the values with those from a scoped enum called StyleAnimationPlayState.
Differential Revision: https://phabricator.services.mozilla.com/D9382
--HG--
extra : moz-landing-system : lando
This pulls a new nom version, which is slightly unfortunate, but I do want some
of the fixes upstream, and it's build-only, so I think it's not a huge deal.
Differential Revision: https://phabricator.services.mozilla.com/D9362
Have embedders send DOM keys to servo and use a strongly typed KeyboardEvent
from the W3C UI Events spec. All keyboard handling now uses the new types.
Introduce a ShortcutMatcher to recognize key bindings. Shortcuts are now
recognized in a uniform way.
Updated the winit port.
Updated webdriver integration.
part of #20331
This cherry-picks part of servo/servo#21881.
Transitions are still broken, but I found these messages more helpful than the
previous ones when diagnosing problems.
This cherry-picks part of servo/servo#20757.
The last caller who used was #14418, which did fix a problem but introduced
multiple. In particular, now transitions don't get expired ever, until they
finish running of course.
That is not ok, given you can have something that the user can trigger to change
the style (hi, :hover, for example), and right now that triggers new
transitions, getting this into a really funny state.
I should give fixing this a shot, but it's non-trivial at all.
This cherry-picks part of servo/servo#20757.
This was done that way just because Servo didn't support the `all` property at
the time.
We should do it this way and optimize it if it's slow. Though I suspect that
most of stuff doesn't actually need to be inherited, my patch at bug 1498943
should make it much faster than what it would otherwise be.
This cherry-picks servo/servo#21946.
This implements the selector(<complex-selector>) syntax for @supports.
See https://github.com/w3c/csswg-drafts/issues/3207 for explainer and
discussion.
Probably would should wait for that to be sorted out to land this, or maybe we
should put it behind a pref to get the code landed and change our
implementation if the discussion there leads to a change.
Differential Revision: https://phabricator.services.mozilla.com/D8864
--HG--
extra : moz-landing-system : lando
This fixes the issue that we should no longer be looking for control characters.
MozReview-Commit-ID: 8k89Aheq3NY
Differential Revision: https://phabricator.services.mozilla.com/D8876
--HG--
extra : moz-landing-system : lando
Move MozLength and MaxLength into generics, and drop the manual implementation
of ToComputedValue.
Depends on D8290
Differential Revision: https://phabricator.services.mozilla.com/D8291
--HG--
extra : moz-landing-system : lando
In order to get the correct computed value of these keywords, we have to
make sure we store the correct computed values in sizing properties in
both inline axis and block axis.
-moz-max-content and -moz-min-content should behave as the property's
initial value in block axis. -moz-fit-content and -moz-available are not
supported in block axis, so we also treat them as initial values.
Differential Revision: https://phabricator.services.mozilla.com/D8290
--HG--
extra : moz-landing-system : lando
I plan to change servo to use all: inherit on its UA sheet.
I hope the patch below should make it good enough performance-wise. And also,
it's probably broken so I don't think it's worth supporting it specially.
Differential Revision: https://phabricator.services.mozilla.com/D8686
--HG--
extra : moz-landing-system : lando
This makes us not allocate useless style structs when you're doing something
like resetting an already-reset property, or inheriting an already-inherited
property.
Seemed simple enough that I think we should do it. In practice we don't even
should pay an extra branch because I expect the compiler to be smart enough and
merge it with the one in the mutate() call.
Differential Revision: https://phabricator.services.mozilla.com/D8685
--HG--
extra : moz-landing-system : lando
This matches the spec, https://drafts.csswg.org/css-values/#angles, which says:
> All <angle> units are compatible, and deg is their canonical unit.
And https://drafts.csswg.org/css-values/#compat, which says:
>When serializing computed values [...], compatible units [...] are converted into a single canonical unit.
And also other implementations (Blink always serializes angles as degrees in
computed style for example).
Also allows us to get rid of quite a bit of code, and makes computed angle value
representation just a number, which is nice.
Differential Revision: https://phabricator.services.mozilla.com/D8619
--HG--
extra : moz-landing-system : lando
I had this around, I couldn't work on more stuff today, but I may as well land
this.
Differential Revision: https://phabricator.services.mozilla.com/D8277
--HG--
extra : moz-landing-system : lando
I had to fix the conversion for BackgroundSize too, hopefully we can
simplify all this using cbindgen in the future instead of CalcValue.
Differential Revision: https://phabricator.services.mozilla.com/D7580
--HG--
extra : moz-landing-system : lando
Specifically:
- nsICSSAnonBoxPseudo --> nsCSSAnonBoxPseudoStaticAtom
- nsICSSPseudoElement --> nsCSSPseudoElementStaticAtom
The `nsI` prefix isn't necessary because these are no longer XPIDL types, and
the `StaticAtom` suffix makes their meaning clearer.
--HG--
extra : rebase_source : b68dd7c73f5036dcd6be4c3700b757441f59f9f2
This saves one word per static atom, per process.
The `nsGkAtoms` change is only a small part of this commit.
In regen_atoms.py:
- There is now only one link name per platform: nsGkAtoms::sAtoms[].
- But there is a new constant per atom, giving the index into
nsGkAtoms::sAtoms[].
- And the `atom!` macro for each atom indexes into nsGkAtoms::sAtoms[] using
the index constant.
- A couple of `*mut` pointers are now `*const`.
Elsewhere, the `(nsStaticAtom*)` casts within the `AppendElement()` calls are
necessary to avoid link errors, presumably due to some template instantiation
wrinkle.
--HG--
extra : rebase_source : 629642e708c8bc6e27d6057beae5f35955fdd837
Specifically, give all the string templates the ''' form, and give them all
`_TEMPLATE` suffixes. This requires slightly changing the newline handling.
--HG--
extra : rebase_source : c89a77bc6d2cbc19cfaf51bbc694c4f3f8be10ba
This commit adds the multi-position gradient color-stops syntax.
GradientItem::parse_comma_separated is extended to attempt to parse
a LengthOrPercent after each color stop. If it succeeds, it appends
an additional color stop with a duplicate color and the specified
position.
This change is only to the parsing, serialization is left unchanged
as per [1].
[1] https://github.com/w3c/csswg-drafts/issues/2714
Differential Revision: https://phabricator.services.mozilla.com/D7380
--HG--
extra : rebase_source : c580e4bd0e4013d19d8418870aa04eb56fa29303
extra : intermediate-source : e85309ff67f47c2ee52944499f6ababe1278fabd
extra : source : 18e81538f6f439a8c02a03f21fff22cc5e6128f4
This commit adds the multi-position gradient color-stops syntax.
GradientItem::parse_comma_separated is extended to attempt to parse
a LengthOrPercent after each color stop. If it succeeds, it appends
an additional color stop with a duplicate color and the specified
position.
This change is only to the parsing, serialization is left unchanged
as per [1].
[1] https://github.com/w3c/csswg-drafts/issues/2714
Differential Revision: https://phabricator.services.mozilla.com/D7380
--HG--
extra : rebase_source : 88f2e9276c626bf4e195f10c5b5713861eeb5760
extra : source : 18e81538f6f439a8c02a03f21fff22cc5e6128f4
This changes the serialization of -moz-image-region to be consistent with that
of clip(), but it's an internal property and the specified value is already
serialized by servo so I think it's the right thing to do.
There are also no internal callers of the getter.
Differential Revision: https://phabricator.services.mozilla.com/D7069
Replace mozilla.widget.use-argb-visuals pref by -moz-gtk-csd-transparent-background media query at browser.css
to draw transparent background. The media query is set by toolkit code and enabled when compatible
window manager is detected.
Differential Revision: https://phabricator.services.mozilla.com/D6658
--HG--
extra : moz-landing-system : lando
This reverts the relevant bits from #21746 so that style and dependencies can
build with stable.
This is important because:
* `selectors` is a published crate.
* Gecko compiles with stable (more or less).
I reviewed that PR under the assumption that the union feature was stable, since
untagged unions are stable since 1.19, but turns out that smallvec uses non-Copy
types in unions, which are still unstable.
This leaves the union feature used on Servo, so that it gets testing, taking
advantage of features being additive.
This cherry-picks servo/servo#21788.
We had a mix of 0.6.2 and 0.6.5 (which is the current release),
this unifies to the latest version. It also enables the union
feature which removes the discriminant, reducing memory usage.
This cherry-picks servo/servo#21746.
According to the new svg 2 spec update (#543), we flip the flag half way for
path interpolation.
Differential Revision: https://phabricator.services.mozilla.com/D6192
--HG--
extra : moz-landing-system : lando
Summary:
The behavior the WG proposed is way more subtle than what that bug implements,
including:
* Implementing two logical overflow longhands.
* Expanding the overflow shorthand to different longhands depending on the
syntax of that.
Meanwhile, Blink hasn't done the swap and will ship the same behavior that we
shipped in Firefox 61 (bug 1453148), that is, overflow-x, then overflow-y.
So I think lacking a clear way forward we should revert this change and preserve
our shipped behavior.
Reviewers: dbaron!
Tags: #secure-revision
Bug #: 1492567
Differential Revision: https://phabricator.services.mozilla.com/D6317
I always see a bunch of time in our profiles in the iterator over the
declarations, this ensures it's not something dumb.
I suspect it's just a bunch of cache misses from walking the rule tree but in
any case this is consistent with the other getters we have and such.
Differential Revision: https://phabricator.services.mozilla.com/D5971
--HG--
extra : moz-landing-system : lando
I think it used to be the case that all PropertyDeclaration variants had a
DeclaredValueOwned<T> inside. But that's no longer the case, so this abstraction
seems less useful now.
Differential Revision: https://phabricator.services.mozilla.com/D5978
--HG--
extra : moz-landing-system : lando
There's no good reason we construct a DeclaredValue as an intermediate step.
Differential Revision: https://phabricator.services.mozilla.com/D5977
--HG--
extra : moz-landing-system : lando
That way I can unify Servo's and Gecko's Cargo.toml files.
This is the only conflict that exists since a while ago, and it causes
continuous manual merges when I sync changes around.
This moves it to ports/geckolib, which works equally well and isn't in the Servo
repo.
Differential Revision: https://phabricator.services.mozilla.com/D5957
--HG--
extra : moz-landing-system : lando
There are a few mentions of nsRuleNode left but they are mostly
historical references so it makes sense to keep them.
Differential Revision: https://phabricator.services.mozilla.com/D5505
We define the standalone types for using derive macro easily and overriding
the behaviors of this traits. This could avoid defining the general
behavior of booleans.
Depends on D4788
Differential Revision: https://phabricator.services.mozilla.com/D4813
--HG--
extra : moz-landing-system : lando
Here, we change the animation type of offset-path as ComputedValue, so
we could do animation on it. Also enable the wpt for offset-path
interpolation. In test_transition_per_property.html, we add some basic tests
ifor offset-path.
ToAnimatedZero for PathCommand will be dropped later.
Because the animations of arcs with mismatched flags are fallen back to
discrete animations, the result of getComputedValue is not normalized in this
case. This makes some wpt failed even though the progress is 100%.
Depends on D4786
Differential Revision: https://phabricator.services.mozilla.com/D4787
--HG--
extra : moz-landing-system : lando
Implement Animate trait for SVGPathData.
The basic idea is: we normalize |this| and |other| svg paths, and then
do interpolation on the normalized svg paths. The normalization is to
convert relative coordinates into absolute coordinates, so we could do
real number interpolation on each path command directly.
In this patch, we also make |clip-path:path()| animatable.
Differential Revision: https://phabricator.services.mozilla.com/D4786
--HG--
extra : moz-landing-system : lando
Looks like this produces sensible results for interpolation with 0, though I'm
not really convinced about the results from, let's say, 1px to 2000px in the
attached test-case, I would've expected a linear interpolation from that to go
through normal length interpolation.
css-transforms-1 says:
> Two transform functions with the same name and the same number of arguments
> are interpolated numerically without a former conversion. The calculated
> value will be of the same transform function type with the same number of
> arguments.
>
> Special rules apply to <matrix()>.
Which is what we do... I was going to file a spec issue but turns out that it's
already addressed in css-transforms-2:
https://drafts.csswg.org/css-transforms-2/#interpolation-of-transform-functions
Which says:
> The transform functions <matrix()>, matrix3d() and perspective() get
> converted into 4x4 matrices first and interpolated as defined in section
> Interpolation of Matrices afterwards.
Differential Revision: https://phabricator.services.mozilla.com/D4942
--HG--
extra : moz-landing-system : lando
Add a preference, layout.css.clip-path-path.enabled, for |clip-path:path()|.
Differential Revision: https://phabricator.services.mozilla.com/D4965
--HG--
extra : moz-landing-system : lando
This also removes one of my FIXMEs from when I was looking at this code.
We don't seem to have a pre-existing test for this feature, sigh. I'll try to
write one if I have cycles for it...
Note that it not applying XBL rules is a feature, given the current state of
affairs. Video controls and such are right now unusable with no styles enabled.
Differential Revision: https://phabricator.services.mozilla.com/D4795
--HG--
extra : moz-landing-system : lando
We always serialize as an atom, which is the previous behavior (though previous
code was using string escaping which I think was not totally sound either...).
Differential Revision: https://phabricator.services.mozilla.com/D4753
--HG--
extra : moz-landing-system : lando
We're wasting 1kb there, which is kind of stupid.
The only advantage of using a bloom filter is that memory usage doesn't increase
even if there's a gazillion attribute selectors and such. But:
* For IDs we already have a bunch of data structures for invalidation and such
which key on the id, so the bloom filter would be a very minor thing.
* For attribute selectors we don't have such a data structure, but if people
used a gazillion attribute selectors we should!
Differential Revision: https://phabricator.services.mozilla.com/D4668
--HG--
extra : moz-landing-system : lando
So that we don't waste a bunch of memory with stuff like <svg:use>. I
plan to shrink AuthorStyles further, but this should help regardless, and isn't
very complex.
Differential Revision: https://phabricator.services.mozilla.com/D4618
--HG--
extra : moz-landing-system : lando
We only care about ascii char for svg path, so we could parse the string
as byte slice.
Differential Revision: https://phabricator.services.mozilla.com/D4168
--HG--
extra : moz-landing-system : lando
For now, |clip-path: path()| is chrome-only, and not for shape-outside,
so we only implement the parser for clip-path. Besides, I didn't put
path() in BasicShape because path() doesn't use the reference box to
resolve the percentage or keywords (i.e. SVG path only accept floating
point or integer number as the css pixel value). Therefore, I add it into
ShapeSource, instead of BasicShape.
Differential Revision: https://phabricator.services.mozilla.com/D3633
SVGPathData will be used by clip-path and offset-path (and/or more on the
properties which support <basic-shape>). Therefore, let's move
SVGPathData out of motion.rs.
Differential Revision: https://phabricator.services.mozilla.com/D3631
Just a minor fix to use cbindgen to avoid the conversion between
basic_shape::FillRule and mozilla::StyleFillRule.
Differential Revision: https://phabricator.services.mozilla.com/D4171
--HG--
extra : moz-landing-system : lando
The implementation of ToCSS for Polygon has some rule, and we could use skip_if
to handle and serialization of fill-rule. However, we should derive ToCSS for
the pair of LengthOrPercentages, so define a new type for it.
Differential Revision: https://phabricator.services.mozilla.com/D4153
--HG--
extra : moz-landing-system : lando
This was consistently faster in the benchmark (even when counters were disabled,
which was slightly suspicious, but...).
Anyway, it's not really much code, most of it is FFI copy-pasta.
Differential Revision: https://phabricator.services.mozilla.com/D3874
Mostly testing that they work, and that they record what we expect them to
record, that is, the actual property that was parsed, and not the properties
that it'd resolve or expand to.
That may be another tricky part for CSSOM, I think style setters would fail an
alias test if implemented with the current setup (we do the property lookup in
C++).
Differential Revision: https://phabricator.services.mozilla.com/D3829
Still not hooked into telemetry, I talked with :janerik and :gfritzsche about
that, but test incoming!
This intentionally doesn't handle CSSOM and such for now, will file followups
for those, though should be trivial.
I want to unify / clean up how we do the use counters and the error reporting
stuff for CSSOM, since the current function call still shows up in profiles,
but that should be a follow-up.
Differential Revision: https://phabricator.services.mozilla.com/D3828
There are a lot of duplicates, so we use macro to refine them.
Depends on D2963
Differential Revision: https://phabricator.services.mozilla.com/D2966
--HG--
extra : moz-landing-system : lando
Define OffsetPath & SVGPathData on the servo-side, and StyleMotion &
StyleSVGPath on the gecko-side. We parse the SVG Path string into a
vector of PathCommand. To build the gfx::Path, we will convert it into
gfx::Path later in a different patch.
The basic flow is:
- Parse SVG Path String into SVGPathData (in Rust).
- Use cbindgen to make sure the layout of PathCommand and StylePathCommand, and then set the Box[PathCommand] into nsTArray<StylePathCommand>.
- Try to convert nsTArray<StylePathCommand> into gfx::Path. (This part will be implemented in a different patch.)
Finally, we use the gfx::Path to create a motion path transform.
The layout implementation is in the later patch.
Depends on D2962
Differential Revision: https://phabricator.services.mozilla.com/D2963
--HG--
extra : moz-landing-system : lando
Define OffsetPath & SVGPathData on the servo-side, and StyleMotion &
StyleSVGPath on the gecko-side. We parse the SVG Path string into a
vector of PathCommand. To build the gfx::Path, we will convert it into
gfx::Path later in a different patch.
The basic flow is:
* Parse SVG Path String into SVGPathData (in Rust).
* Use cbindgen to make sure the layout of PathCommand and StylePathCommand,
and then set the Box[PathCommand] into nsTArray<StylePathCommand>.
* Try to convert nsTArray<StylePathCommand> into gfx::Path. (This part
will be implemented in a different patch.)
Finally, we use the gfx::Path to create a motion path transform.
The layout implementation is in the later patch.
Differential Revision: https://phabricator.services.mozilla.com/D2963
Port `width`, and also add the `scan` media feature so I don't need to add
ugliness just to workaround the unused keyword_evaluator macro.
This is not part of the Gecko build in any case.
To support that, this patch also does the following.
- Removes the insert(), remove() and might_contain() methods, because they are
specialized versions of insert_hash(), remove_hash(), and
might_contain_hash(), and they are only used by tests within this file.
- Moves hash() from the top level into create_and_insert_some_stuff().
- Changes create_and_insert_some_stuff() so that instead of hashing consecutive
integers, it instead hashes stringified consecutive integers, which matches
real usage a little better.
- Raises the false_positives limit a little to account for the above changes.
--HG--
extra : rebase_source : f839ff86986d5c8a17506a808ba676f69c7ef407
Now that :-moz-system-metric is gone, there's no real reason for the atoms to
be separate.
Differential Revision: https://phabricator.services.mozilla.com/D3497
--HG--
extra : moz-landing-system : lando
This moves most of the code to be Rust, except potentially some evaluator
functions, and allows to unblock the use case from any-hover / any-pointer and
remove nsMediaFeatures.
Differential Revision: https://phabricator.services.mozilla.com/D2976
We match with AllLinksVisitedAndUnvisited for style invalidation, and we already
do a subtree restyle because :visited matching doesn't depend on the actual
element state.
So all this stuff is just not needed. The comment points to the attribute tests
in bug 1328509, but those still trivially pass with this change.
I think this was unneeded since I introduced AllLinksVisitedAndUnvisited, or
maybe since https://github.com/servo/servo/pull/19520. In any case it doesn't
really matter, and I already had done this cleanup in my WIP patches for
bug 1406622, but I guess this is a slightly more suitable place to land them :)
Differential Revision: https://phabricator.services.mozilla.com/D3305
We force a repaint from ContentStateChangedInternal if visited links are
disabled, and that's observable. Let's cut it off as early as we can to avoid
timing attacks even when :visited is disabled.
Differential Revision: https://phabricator.services.mozilla.com/D3304
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
