Hashbrown grows a lot sometimes making us waste a lot of memory. Shrink
some of these maps after CascadeData rebuild / stylesheet collection
invalidation.
Differential Revision: https://phabricator.services.mozilla.com/D134716
The bitfield approach worked when the layer order was in pre-order, but
the spec was changed to make it work like post-order and I don't think
there's a way to keep it working like that, so keep the layer order in a
separate data structure that we look up when going from Rule to
ApplicableDeclarationBlock.
This is just a vector index operation so hopefully shouldn't be too bad.
This patch intentionally regresses @keyframe handling to some extent,
since we need a bit more complicated approach and it seemed worth
implementing in a separate patch.
Depends on D129380
Differential Revision: https://phabricator.services.mozilla.com/D129381
See the discussion here: https://twitter.com/Rich_Harris/status/1433153204678799365
This should make attribute selectors roughly as fast as class selectors.
I think it's worth trying and see if perf bots complain on
micro-benchmarks and stylebench and such.
I made attributes more specific than local names, but less specific than
classes, which I think makes sense. When doing something like
foo[data-bar], filtering by data-bar seems likely to yield less elements
than filtering by foo.
While at it, remove the bloom filter pref since we shipped it in
bug 1704551 for 87 and we haven't heard complaints.
Differential Revision: https://phabricator.services.mozilla.com/D124383
This should be both a memory and speed win for pages using a lot of
Shadow DOM.
In order to make the cache properly work we need to start keying media query
results on the actual StyleSheetContents, as that's what we share on Gecko, but
that should all be fine.
Differential Revision: https://phabricator.services.mozilla.com/D107266
This fixes performance issues when there's a massive number of rules and
a massive DOM.
Instead of going through all rules for all the DOM, we now look stuff in
hashmaps.
Differential Revision: https://phabricator.services.mozilla.com/D83506
This way, something like:
*:where(.foo, .bar)
Will end up twice on the selector map, just as if you would've written
.foo, .bar.
But we're a bit careful to not be wasteful, so:
.foo:where(div, span)
Will still end up using the .foo bucket.
It needs a bit of borrow-checker gymnastics to avoid cloning the entry
in the common path. It's a bit gross but not too terrible I think.
Differential Revision: https://phabricator.services.mozilla.com/D71457
This implements the easy / straight-forward parts of the :where / :is
selectors.
The biggest missing piece is to handle properly invalidation when there
are combinators present inside the :where. That's the hard part of this,
actually.
But this is probably worth landing in the interim. This fixes some of
the visitors that were easy to fix.
Differential Revision: https://phabricator.services.mozilla.com/D70788
I still haven't implemented each_part(), so this will do nothing yet.
The cascade order stuff is fishy, I know, and I'll fix in a followup if it's
fine with you. I moved the sorting of the rules to rule_collector, since it
seemed to me it was better that way that duplicating the code, and those
SelectorMap functions only have a single caller anyway.
Differential Revision: https://phabricator.services.mozilla.com/D32647
: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
After bug 1470163 we have some nasty selectors from mathml.css in every page.
We only want to match them against MathML elements.
This patch brings the global revalidation selectors from 14 to 2 in about:blank.
Also halves the ones from XUL documents.
Bug: 1374017
Reviewed-by: heycam
MozReview-Commit-ID: nOVyknNcVm
No cleaner ideas right now that carrying that counter around... Maybe a custom
type may be cleaner?
This makes ApplicableDeclarationBlock a bit bigger. I could probably try to make
the counter a 4 / 5-bit number or something and pack the counter there in the
SourceOrderAndCascadeLevel somehow...
But doesn't seem really worth the churn, and can be done as a followup in any
case. Let me know if you want to block on that.
Bug: 1454162
Reviewed-by: heycam
MozReview-Commit-ID: 1LdW9S4xA6f
We could invalidate in a slightly more fine-grained way, but I don't think it's
worth the churn vs. keeping the special-cases minimal.
Bug: 1452640
Reviewed-by: xidorn
MozReview-Commit-ID: 5DkQrgwg9GW
All TElement's implement Copy, and are just pointers, so the double indirection
is stupid.
I'm going to try to see if removing this double-indirection fixes some
selector-matching performance, and this is a trivial pre-requisite while I wait
for Talos results.
This function is only ever used with one type. This gets rid of the
only use of the `smallvec::VecLike` trait, which we may want to
deprecate. (If we do need to make this function generic in the future,
we can do it using standard traits instead.)
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`.
This patch makes the MallocSizeOf stuff in Stylo work more like the HeapSizeOf
stuff already in Servo, except better. In particular, it adds deriving support
for MallocSizeOf, which will make it easier to improve coverage.
The patch does the following.
- Combines servo/components/style/stylesheets/memory.rs and the heapsize crate
into a new crate, malloc_size_of.
- Forks the heapsize_derive crate, calling it malloc_size_of, so that
MallocSizeOf can be derived.
- Both the new crates have MIT/Apache licenses, like heapsize, in case they are
incorporated into heapsize in the future.
- Renames the methods within MallocSizeOf and the related traits so they are
more concise.
- Removes MallocSizeOfWithGuard.
- Adds `derive(MallocSizeOf)` to a lot of types, in some cases replacing an
equivalent or almost-equivalent hand-written implementation.
- Adds stuff so that Rc/Arc can be handled properly.
