Produce JS-visible string results as UTF-16 at their source, including
numeric formatting, BigInt and BigFraction formatting, URI encoding,
console formatting, parser errors, regular expression errors, Intl and
Temporal records, LibUnicode locale boundaries, and LibWeb bindings.
Handle fractional radix formatting through the UTF-16 builder view.
Thread UTF-16 string input through JSON, script parsing, Date parsing,
Intl option parsing, Temporal parsing, and the helper library boundaries
that feed those parsers. Preserve ASCII fast paths where the source data
is known to be ASCII.
Move the remaining LibJS primitive string users to UTF-16 views and
strings. Remove the primitive string UTF-8 accessors and byte-string
coercion paths so new callers cannot rely on the old storage model.
Keep primitive string storage in Utf16String and remove the UTF-8
storage path from PrimitiveString. ASCII strings still use compact
Utf16String ASCII storage, while UTF-16 becomes the only owned
representation.
Port more UTF-16 string construction sites to Utf16StringBuilder.
This covers JSON serialization, URI escaping, RegExp replacement, date
and Temporal formatting, Uint8Array hex conversion, and stack string
formatting. Keep byte-oriented debug, bytecode, parser, and print
plumbing on StringBuilder.
Add a checked JS string length sum helper and use it for accumulation
paths that append JS-observable string pieces.
Enable -Wexit-time-destructors for all in-tree library targets and
update process-lifetime library statics so they no longer register
exit-time destructors. Long-lived caches, lookup tables, singleton
registries, and generated constants now use NeverDestroyed or leaked
references where the data is intended to live until process exit.
Update LibWeb, LibLine, and the binding generators so regenerated
sources follow the same rule instead of reintroducing destructed
statics.
This better describes what the method returns and avoids the possible
confusion caused by the mismatch in behavior between
`Value::is_array()` and `Value::as_array()`.
Property lookup cache entries previously used GC::Weak<T> for shape,
prototype, and prototype_chain_validity pointers. Each GC::Weak
requires a ref-counted WeakImpl allocation and an extra indirection
on every access.
Replace these with GC::RawPtr<T> and make Executable a WeakContainer
so the GC can clear stale pointers during sweep via remove_dead_cells.
For static PropertyLookupCache instances (used throughout the runtime
for well-known property lookups), introduce StaticPropertyLookupCache
which registers itself in a global list that also gets swept.
Now that inline cache entries use GC::RawPtr instead of GC::Weak,
we can compare shape/prototype pointers directly without going
through the WeakImpl indirection. This removes one dependent load
from each IC check in GetById, PutById, GetLength, GetGlobal, and
SetGlobal handlers.
In the benchmark added here, fmt's dragonbox is ~3x faster than our own
Ryu implementation (1197ms for dragonbox vs. 3435ms for Ryu).
Daniel Lemire recently published an article about these algorithms:
https://lemire.me/blog/2026/02/01/converting-floats-to-strings-quickly/
In this article, fmt's dragonbox implementation is actually one of the
slower ones (with the caveat that some comments note that the article is
a bit out-of-date). I've gone with fmt here because:
1. It has a readily available recent version on vcpkg.
2. It provides the methods we need to actually convert a floating point
to decimal exponential form.
3. There is an ongoing effort to replace dragonbox with a new algorithm,
zmij, which promises to be faster.
4. It is one of the only users of AK/UFixedBigInt, so we can potentially
remove that as well soon.
5. Bringing in fmt opens the door to replacing a bunch of AK::format
facilities with fmt as well.
This patch improves JSON.stringify performance through three changes:
1. Use a single StringBuilder for the entire operation instead of
building up intermediate strings and concatenating them.
2. Format numbers directly into the StringBuilder via a new public
number_to_string(StringBuilder&, ...) overload, avoiding temporary
String allocations.
3. Track indentation as a depth counter instead of repeatedly
concatenating the gap string.
This is to make it much easier when trying to pull out an object
of a specific type from a JS::Value, instead of needing to do the
repetitive checking that it is an object, getting that object,
checking that object is of a specific type, then casting to that
type.
The bytecode interpreter already implements these fast paths, usually
more efficiently as well.
Turns out this stuff was getting inlined into the interpreter, so we
had many fast paths twice(!)
When testing with JS_BYTECODE_DEBUG macro defined or using the All_Debug
profile, internal objects were not known in Value class to the function
Value::to_string_without_side_effects leading to a VERIFICATION FAILED
when running the test-js or test-web programs.
Internal objects are known in the Value class as cells, and do not have
a dedicated tag to identify them. Internal objects are detected using
is_cell function call, but only after all other types have been
checked as other types of non-internal objects can also be cells.
Both the String and Utf16String version of the function were updated.
We can use caching in a million more places. This is just me running JS
benchmarks and looking at which get() call sites were hot and putting
caches there.
Lots of nice speedups all over the place, some examples:
1.19x speedup on Octane/raytrace.js
1.13x speedup on Octane/earley-boyer.js
1.12x speedup on Kraken/ai-astar.js
1.10x speedup on Octane/box2d.js
1.08x speedup on Octane/gbemu.js
1.05x speedup on Octane/regexp.js
To speed up property access, callers of get() can now provide a lookup
cache like so:
static Bytecode::PropertyLookupCache cache;
auto value = TRY(object.get(property, cache));
Note that the cache has to be `static` or it won't make sense!
This basically brings the inline caches from our bytecode VM straight
into C++ land, allowing us to gain serious performance improvements.
The implementation shares code with the GetById bytecode instruction.
This makes the instanceof operator signficantly faster by avoiding a
generic function call to @@hasInstance unless it has been overridden.
1.15x speed-up on Octane/earley-boyer.js
We are often forced to convert numbers to strings inside LibJS, e.g when
iterating over the property names of an array, but it's also just a very
common operation in general.
This patch adds a 1000-entry string cache for the numbers 0-999 since
those appear to be by far the most common ones we convert.
Adds inline implementation for the most common case when `Value` is
already an object.
1.47x improvement on the following benchmark:
```js
const o = {};
for (let i = 0; i < 10_000_000; i++) {
o.a = 1;
o.b = 2;
o.c = 3;
}
```
This has quite a lot of fall out. But the majority of it is just type or
UDL substitution, where the changes just fall through to other function
calls.
By changing property key storage to UTF-16, the main affected areas are:
* NativeFunction names must now be UTF-16
* Bytecode identifiers must now be UTF-16
* Module/binding names must now be UTF-16
Our floating point number parser was based on the fast_float library:
https://github.com/fastfloat/fast_float
However, our implementation only supports 8-bit characters. To support
UTF-16, we will need to be able to convert char16_t-based strings to
numbers as well. This works out-of-the-box with fast_float.
We can also use fast_float for integer parsing.
Having it as a method instead of a free function is necessary for the
next commits and generally allows for optimizations that require deeper
access into the `UnsignedBigInteger` / `SignedBigInteger`.
Also restrict the exponent to 32 bits to avoid huge memory allocations.
- Avoids unnecessary conversions between StringOrSymbol and PropertyKey
on the hot path of property access.
- Simplifies the code by removing StringOrSymbol and using PropertyKey
directly. There was no reason to have a separate StringOrSymbol type
representing the same data as PropertyKey, just with the index key
stored as a string.
PropertyKey has been updated to use a tagged pointer instead of a
Variant, so it still occupies 8 bytes, same as StringOrSymbol.
12% improvement on JetStream/gcc-loops.cpp.js
12% improvement on MicroBench/object-assign.js
7% improvement on MicroBench/object-keys.js
This is very frequently returned by Object.prototype.toString(),
so we benefit from caching it instead of recreating it every time.
Takes Speedometer2.1/EmberJS-Debug-TodoMVC from ~4000ms to ~3700ms
on my M3 MacBook Pro.
If we have two PrimitiveString objects that are both backed by UTF-16
data, we don't have to convert them to UTF-8 for equality checking.
Just compare the underlying UTF-16 data. :^)
FJCVTZS (Floating-point Javascript Convert to Signed fixed-point,
rounding toward Zero) does exactly what we need for ToInt32 in the
JavaScript specification.
This isn't world-changing, but it does give a ~2% boost on compute-
heavy benchmarks like JetStream, so we should obviously use it.
The fast path of to_i32() can be neatly inlined everywhere, and we still
have to_i32_slow_case() for non-trivial conversions.
For to_u32(), it really can just be implemented as a static cast to i32!
The special empty value (that we use for array holes, Optional<Value>
when empty and a few other other placeholder/sentinel tasks) still
exists, but you now create one via JS::js_special_empty_value() and
check for it with Value::is_special_empty_value().
The main idea here is to make it very unlikely to accidentally create an
unexpected special empty value.
PrimitiveString is now internally either UTF-8, UTF-16, or both.
We no longer convert them to/from ByteString anywhere, nor does VM have
a ByteString cache.
