This will ultimately make it simpler to update crate dependencies and
reduce duplicate when specifying requirements. Generally, this change
does not touch dependencies that are only used by a single crate. We
could consider moving them to workspace dependencies in the future.
Release selectors 0.24.0 and servo_arc 0.2.0
This patch extends #29361, bumping selectors to 0.24.0 and servo_arc to 0.2.0.
We’ll need to publish them too, in order to close#29105.
---
<!-- 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#29105 (GitHub issue number if applicable)
<!-- Either: -->
- [ ] There are tests for these changes OR
- [x] These changes do not require tests because there are no functional changes
The existing code wasn't sound, as CSSOM objects also needed to go away before
the shared memory goes away (as they keep references to them).
This is sound assuming no presence of reference cycles introduced by CSSOM.
We may want to live with this and rely on chrome code not writing cycles like
this with UA stylesheet DOM objects.
We could explicitly drop all potentially-static objects... That seems pretty
error prone though.
Or we could also just leak the shared memory buffer, is there any reason why we
may not want to do that?
Differential Revision: https://phabricator.services.mozilla.com/D51870
This needs https://github.com/eqrion/cbindgen/pull/362, but I expect it to be
uncontroversial. I'll add a patch to this bug when it's merged to update it.
cbindgen historically didn't include these, but it turns out to be pretty useful
to generate constants for the style crate (since the binding crate is
`servo/ports/geckolib`).
An alternative is to get a completely different cbindgen-generated header for
these, but that seems a bit wasteful. This generates the constants with the
Style prefix (so we'll get `StyleMAX_GRID_LINE` for example), which is very
ugly. But we probably want to eventually stop using the Style prefix and use a
namespace instead, plus it's trivial to do `auto kMaxLine = StyleMAX_GRID_LINE`,
for example, so it's probably not a huge deal.
Another alternative would be to use associated consts, which _are_ generated by
cbindgen. Something like:
```
struct GridConstants([u8; 0]);
impl GridConstants {
const MAX_GRID_LINE: i32 = 10000;
}
```
Which would yield something like:
```
static const int32 StyleGridConstants_MAX_GRID_LINE = 10000;
```
I'm not sure if you find it preferrable, but I'm also happy to change it in a
follow-up to use this.
We need to fix a few manual C++ function signature definitions to match the C++
declaration.
Differential Revision: https://phabricator.services.mozilla.com/D35197
I need to profile this a bit more, but talos was pretty happy about this, and it
solves the known performance issues here such as the test-case from bug 1483963
for example. This also gets rid of a bunch of unsafe code which is nice.
This still keeps the same GC scheme, removing the key from the hashmap when
needed. I kept those as release assertions, but should probably be turned into
debug-only assertions.
Differential Revision: https://phabricator.services.mozilla.com/D6801
We cannot compile with just feature(gecko + debug_assertions), since that's how
debug rusttests get compiled and they don't have the refcount logging stuff.
We were getting away with it for the pre-existing usage of the style crate,
because it wasn't used during any test and presumably the linker didn't
complain. But servo_arc is definitely used in tests.
Differential Revision: https://phabricator.services.mozilla.com/D32691
This doesn't clean up as much as a whole, but it's a step in the right
direction. In particular, it allows us to start using simple bindings for:
* Filters
* Shapes and images, almost. Need to:
* Get rid of the complex -moz- gradient parsing (let
layout.css.simple-moz-gradient.enabled get to release).
* Counters, almost. Need to:
* Share the Attr representation with Gecko, by not using Option<>.
* Just another variant should be enough (ContentItem::{Attr,Prefixedattr},
maybe).
Which in turn allows us to remove a whole lot of bindings in followups to this.
The setup changes a bit. This also removes the double pointer I complained about
while reviewing the shared UA sheet patches. The old setup is:
```
SpecifiedUrl
* CssUrl
* Arc<CssUrlData>
* String
* UrlExtraData
* UrlValueSource
* Arc<CssUrlData>
* load id
* resolved uri
* CORS mode.
* ...
```
The new one removes the double reference to the url data via URLValue, and looks
like:
```
SpecifiedUrl
* CssUrl
* Arc<CssUrlData>
* String
* UrlExtraData
* CorsMode
* LoadData
* load id
* resolved URI
```
The LoadData is the only mutable bit that C++ can change, and is not used from
Rust. Ideally, in the future, we could just use rust-url to resolve the URL
after parsing or something, and make it all immutable. Maybe.
I've verified that this approach still works with the UA sheet patches (via the
LoadDataSource::Lazy).
The reordering of mWillChange is to avoid nsStyleDisplay from going over the
size limit. We want to split it up anyway in bug 1552587, but mBinding gains a
tag member, which means that we were having a bit of extra padding.
One thing I want to explore is to see if we can abuse rustc's non-zero
optimizations to predict the layout from C++, but that's something to explore at
some other point in time and with a lot of care and help from Michael (who sits
next to me and works on rustc ;)).
Differential Revision: https://phabricator.services.mozilla.com/D31742
As I said over bug 1549593, the eventual goal is to use ArcSlice in all
inherited properties. But this seemed like a good first candidate that doesn't
require me to move around a lot more code, since we were already using cbindgen
for the path commands.
Differential Revision: https://phabricator.services.mozilla.com/D30134
We could make the header PhantomData or something, but then we wouldn't be able
to bind to C++, since C++ doesn't have ZSTs. So add a canary instead to add a
runtime check of stuff being sane.
Differential Revision: https://phabricator.services.mozilla.com/D30133
The change from [T; 1] to [T; 0] shouldn't change behavior since they have the
right alignment and we never poke at that particular array, but feels more
correct to avoid creating types that point to uninitialized data or outside of
their allocation, now that we allow empty slices.
Differential Revision: https://phabricator.services.mozilla.com/D30352
Servo currently uses `heapsize`, but Stylo/Gecko use `malloc_size_of`.
`malloc_size_of` is better -- it handles various cases that `heapsize` does not
-- so this patch changes Servo to use `malloc_size_of`.
This patch makes the following changes to the `malloc_size_of` crate.
- Adds `MallocSizeOf` trait implementations for numerous types, some built-in
(e.g. `VecDeque`), some external and Servo-only (e.g. `string_cache`).
- Makes `enclosing_size_of_op` optional, because vanilla jemalloc doesn't
support that operation.
- For `HashSet`/`HashMap`, falls back to a computed estimate when
`enclosing_size_of_op` isn't available.
- Adds an extern "C" `malloc_size_of` function that does the actual heap
measurement; this is based on the same functions from the `heapsize` crate.
This patch makes the following changes elsewhere.
- Converts all the uses of `heapsize` to instead use `malloc_size_of`.
- Disables the "heapsize"/"heap_size" feature for the external crates that
provide it.
- Removes the `HeapSizeOf` implementation from `hashglobe`.
- Adds `ignore` annotations to a few `Rc`/`Arc`, because `malloc_size_of`
doesn't derive those types, unlike `heapsize`.
Fixes warnings from rust-lang/rust#44229 when `--enable-commonmark` is
passed to rustdoc.
This is mostly a global find-and-replace for bare URIs on lines by
themselves in doc comments.
