Object specialization #419

HalidOdat · 2020-05-25T21:05:26Z

It changes the following:

Specializes objects
Implements Hash trait for Value
Implements PartialEq trait based on the SameValueZero algorithm for Value (required by Hash trait)
Implements Eq trait for Value (required by Hash trait)
Fixed a bug when comparing object pointers in same_value_non_number
Deprecated ObjectInternalMethods
Fixed Symbol primitive
Added construct_type_error and construct_range_error
Made Number objects constructor and methods more specification compliant.
Abstracted Object name and function object length.
Changed init functions to return (&str, Value)

The problem with the current implementation is that we store the data in an internal slot. The internal slots are stored as a HashMap, which means slow reads and writes. Additionally they are stored as a Value which means we need additional checks to unwrap the Value into a primitive (example f64, String) to access the data. Furthermore we are restricted as to what type of data we can hold, for example if we wanted to implement Set object with native type HashSet, we can not do so in the internal slots as they are specified in the standard (it's only an implementation detail).
This PR aims to allow use of non-Value types in objects as well as removing the need for internals slots for data storage (NumberData, StringData, etc), removing redundant checks for data access. It should increase performance greatly for object access and a bit for object creation. (I'm not sure the benchmarks will show it though, since we do not have benchmarks for objects).

Specialized objects:

github-actions · 2020-05-26T11:52:39Z

Benchmark for `3d93be3`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	428.5±21.73µs	423.8±18.64µs	101%
Expression (Lexer)	2.0±0.10µs	2.1±0.11µs	96%
Expression (Parser)	4.8±0.25µs	4.7±0.24µs	103%
Fibonacci (Execution)	2.8±0.12ms	2.7±0.12ms	106%
For loop (Execution)	450.7±22.68µs	435.2±21.77µs	104%
For loop (Lexer)	5.2±0.25µs	5.3±0.25µs	98%
For loop (Parser)	13.5±0.75µs	13.3±0.74µs	101%
Hello World (Lexer)	970.7±45.03ns	1001.7±47.19ns	97%
Hello World (Parser)	2.2±0.14µs	2.2±0.13µs	103%
Symbols (Execution)	466.5±20.48µs	438.8±19.30µs	106%
	undefined	undefined	100%

github-actions · 2020-05-26T17:20:43Z

Benchmark for `36cc48e`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	446.9±14.88µs	449.5±19.05µs	99%
Expression (Lexer)	2.2±0.09µs	2.1±0.13µs	101%
Expression (Parser)	5.1±0.15µs	5.2±0.28µs	97%
Fibonacci (Execution)	2.7±0.13ms	2.7±0.14ms	99%
For loop (Execution)	462.9±15.70µs	466.7±26.44µs	99%
For loop (Lexer)	5.6±0.20µs	5.6±0.24µs	100%
For loop (Parser)	14.0±0.67µs	14.2±0.66µs	99%
Hello World (Lexer)	1013.2±39.76ns	1008.5±33.04ns	100%
Hello World (Parser)	2.3±0.12µs	2.4±0.10µs	99%
Symbols (Execution)	483.6±23.77µs	467.9±16.13µs	103%
	undefined	undefined	100%

HalidOdat · 2020-05-26T18:25:40Z

@Razican and @jasonwilliams I'm having a bit of trouble with the garbage collector.
I'm getting this error when construction/calling Array:

thread 'main' panicked at 'GcCell<T> already borrowed', <::std::macros::panic macros>:5:6

Any help would be useful

github-actions · 2020-05-26T18:31:19Z

Benchmark for `b52be12`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	418.2±25.13µs	428.6±25.85µs	98%
Expression (Lexer)	2.1±0.13µs	2.1±0.10µs	100%
Expression (Parser)	5.3±0.17µs	5.1±0.95µs	103%
Fibonacci (Execution)	2.8±0.12ms	2.7±0.14ms	101%
For loop (Execution)	454.0±27.35µs	447.2±31.50µs	102%
For loop (Lexer)	5.8±0.22µs	5.5±0.29µs	106%
For loop (Parser)	15.3±0.45µs	14.7±0.90µs	105%
Hello World (Lexer)	995.4±58.92ns	1030.7±68.24ns	96%
Hello World (Parser)	2.4±0.11µs	2.3±0.17µs	108%
Symbols (Execution)	457.7±33.26µs	454.7±31.08µs	101%
	undefined	undefined	100%

Razican · 2020-05-26T18:43:15Z

@Razican and @jasonwilliams I'm having a bit of trouble with the garbage collector.
I'm getting this error when construction/calling Array:
thread 'main' panicked at 'GcCell<T> already borrowed', <::std::macros::panic macros>:5:6
Any help would be useful

I'm not very used to deal with the Gc, but this seems to be that two calls to borrow() were made one after the other without a clone() or a scope ending in the middle.

github-actions · 2020-05-27T09:49:38Z

Benchmark for `b1ff14c`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	430.4±10.16µs	439.4±10.91µs	98%
Expression (Lexer)	1904.4±109.31ns	2.1±0.08µs	90.99999999999999%
Expression (Parser)	5.0±0.16µs	4.9±0.31µs	102%
Fibonacci (Execution)	2.6±0.08ms	2.6±0.09ms	98%
For loop (Execution)	448.2±18.54µs	407.6±19.47µs	110.00000000000001%
For loop (Lexer)	5.4±0.17µs	5.1±0.24µs	106%
For loop (Parser)	13.5±0.30µs	13.1±0.64µs	103%
Hello World (Lexer)	991.5±12.53ns	963.3±44.15ns	103%
Hello World (Parser)	2.3±0.09µs	2.2±0.09µs	102%
Symbols (Execution)	457.3±12.72µs	458.6±10.63µs	100%
	undefined	undefined	100%

boa/src/builtins/value/mod.rs

HalidOdat · 2020-05-27T13:38:57Z

I did some benchmarks on Object creation and access

I think the other execution benchmarks are wrong, we use exec function to benchmark execution, but exec does not just execute the string, it lexes, parsers, creates the interpreter and initializes/creates the realm and then it executes the code, aren't we polluting the result of the benchmark? Shouldn't we isolate what were executing as much as possible, right?
With exec function, I saw a ~4%, when I isolated the execution I saw a ~20% increase for object creation/access.

You can check the benchmark here

What do you think @Razican and @jasonwilliams ?

Razican · 2020-05-27T14:08:25Z

I did some benchmarks on Object creation and access

I think the other execution benchmarks are wrong, we use exec function to benchmark execution, but exec does not just execute the string, it lexes, parsers, creates the interpreter and initializes/creates the realm and then it executes the code, aren't we polluting the result of the benchmark? Shouldn't we isolate what were executing as much as possible, right?
With exec function, I saw a ~4%, when I isolated the execution I saw a ~20% increase for object creation/access.

You can check the benchmark here

What do you think @Razican and @jasonwilliams ?

@jasonwilliams noticed that most of the execution time is spent on Realm creation, so I think we should isolate the actual execution, yes, in order to see if we improve or worsen there.

jasonwilliams · 2020-05-28T07:51:25Z

I think we need to make some benchmark issues and get them in before this so we can have some before/after

github-actions · 2020-05-28T11:04:47Z

Benchmark for `d0b4745`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	455.5±16.60µs	531.8±70.41µs	83%
Expression (Lexer)	2.2±0.09µs	2.3±0.12µs	97%
Expression (Parser)	5.6±0.45µs	5.3±0.17µs	106%
Fibonacci (Execution)	3.0±0.10ms	3.1±0.12ms	97%
For loop (Execution)	500.8±29.44µs	472.2±26.69µs	106%
For loop (Lexer)	6.0±0.25µs	5.6±0.18µs	109.00000000000001%
For loop (Parser)	14.3±0.40µs	14.5±0.69µs	99%
Hello World (Lexer)	1036.1±33.18ns	1036.1±32.72ns	100%
Hello World (Parser)	2.3±0.11µs	2.5±0.19µs	94%
Symbols (Execution)	480.1±12.42µs	483.5±19.79µs	99%
	undefined	undefined	100%

github-actions · 2020-05-28T21:12:10Z

Benchmark for `345bcc1`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	437.7±12.43µs	443.0±20.56µs	99%
Expression (Lexer)	2.1±0.10µs	2.2±0.09µs	99%
Expression (Parser)	5.0±0.21µs	5.0±0.18µs	99%
Fibonacci (Execution)	2.8±0.11ms	2.8±0.11ms	100%
For loop (Execution)	454.5±15.66µs	464.9±16.46µs	98%
For loop (Lexer)	5.5±0.17µs	5.6±0.29µs	98%
For loop (Parser)	13.9±0.81µs	14.3±0.70µs	97%
Hello World (Lexer)	1020.2±45.76ns	1029.9±63.53ns	99%
Hello World (Parser)	2.3±0.09µs	2.4±0.10µs	95%
Symbols (Execution)	473.8±21.90µs	486.5±25.47µs	97%
	undefined	undefined	100%

github-actions · 2020-05-29T02:50:13Z

Benchmark for `38d0eb1`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	427.6±8.26µs	440.7±9.92µs	97%
Expression (Lexer)	2.1±0.03µs	2.1±0.05µs	98%
Expression (Parser)	5.1±0.11µs	5.1±0.07µs	100%
Fibonacci (Execution)	2.6±0.06ms	2.7±0.05ms	100%
For loop (Execution)	445.0±7.99µs	450.1±8.19µs	99%
For loop (Lexer)	5.5±0.12µs	5.5±0.09µs	100%
For loop (Parser)	13.7±0.47µs	14.0±0.69µs	98%
Hello World (Lexer)	996.4±29.51ns	993.8±11.09ns	100%
Hello World (Parser)	2.3±0.04µs	2.3±0.04µs	100%
Symbols (Execution)	453.7±8.68µs	463.8±14.34µs	98%
	undefined	undefined	100%

github-actions · 2020-05-30T13:06:39Z

Benchmark for `494d109`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Create Realm	431.3±12.20µs	431.1±9.46µs	100%
Expression (Lexer)	2.1±0.06µs	2.1±0.04µs	102%
Expression (Parser)	5.0±0.17µs	5.1±0.13µs	99%
Fibonacci (Execution)	2.6±0.06ms	2.6±0.06ms	100%
For loop (Execution)	449.3±13.71µs	458.7±11.23µs	98%
For loop (Lexer)	5.6±0.12µs	5.4±0.21µs	104%
For loop (Parser)	13.6±0.41µs	13.6±0.34µs	100%
Hello World (Lexer)	1018.5±31.89ns	973.5±31.76ns	105%
Hello World (Parser)	2.2±0.05µs	2.2±0.07µs	100%
Symbols (Execution)	462.0±10.12µs	457.6±14.79µs	101%
	undefined	undefined	100%

Razican

OK, an extra bunch of comments. I think one of the best things we could do is using an Object directly instead of a Value as the global object when initializing the Realm.

boa/src/builtins/array/mod.rs

boa/src/builtins/boolean/mod.rs

boa/src/builtins/error/type.rs

Razican · 2020-06-14T11:54:01Z

boa/src/builtins/function/mod.rs

+
+    prototype
+        .as_object_mut()
+        .unwrap()


This has not been changed, right?

Razican · 2020-06-14T11:54:18Z

boa/src/builtins/function/mod.rs

-    parent.set_field(Value::from(name_copy), new_func_obj);
+    parent
+        .as_object_mut()
+        .unwrap()


This was not changed, right?

boa/src/exec/mod.rs

Razican · 2020-06-14T12:25:13Z

boa/src/exec/mod.rs

+            self.throw_type_error("cannot convert null or undefined to Object")?;
+            unreachable!();


Should we do this in this PR or in a different one?

boa/src/exec/mod.rs

github-actions · 2020-06-14T16:40:26Z

Benchmark for `31c4c6c`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Array access (Execution)	23.7±0.92µs	35.3±1.00µs	-32.86%
Array creation (Execution)	6.4±0.22ms	14.8±0.51ms	-56.76%
Array pop (Execution)	2.4±0.08ms	8.7±0.15ms	-72.41%
Create Realm	164.9±6.89µs	468.0±21.22µs	-64.76%
Dynamic Object Property Access (Execution)	11.2±0.37µs	16.4±0.46µs	-31.71%
Expression (Lexer)	1939.1±59.72ns	2.0±0.07µs	-3.05%
Expression (Parser)	4.7±0.23µs	4.6±0.14µs	+2.17%
Fibonacci (Execution)	1928.0±58.42µs	1987.4±64.63µs	-2.99%
For loop (Execution)	54.4±1.93µs	62.9±4.31µs	-13.51%
For loop (Lexer)	5.1±0.22µs	5.0±0.32µs	+2.00%
For loop (Parser)	12.7±0.48µs	12.4±0.56µs	+2.42%
Hello World (Lexer)	873.0±25.60ns	898.0±28.07ns	-2.78%
Hello World (Parser)	2.1±0.09µs	2.1±0.09µs	0.00%
Long file (Parser)	5.8±0.12ms	5.7±0.16ms	+1.75%
Object Creation (Execution)	10.4±0.43µs	14.6±0.32µs	-28.77%
RegExp (Execution)	74.0±4.32µs	80.9±2.42µs	-8.53%
RegExp Literal (Execution)	77.5±3.34µs	90.0±3.39µs	-13.89%
RegExp Literal Creation (Execution)	72.5±3.28µs	81.6±3.01µs	-11.15%
Static Object Property Access (Execution)	10.5±0.49µs	15.4±0.33µs	-31.82%
String comparison (Execution)	10.4±0.23µs	18.8±0.41µs	-44.68%
String concatenation (Execution)	8.7±0.35µs	15.0±0.44µs	-42.00%
String copy (Execution)	7.1±0.27µs	12.5±0.42µs	-43.20%
Symbols (Execution)	8.0±0.26µs	16.9±0.58µs	-52.66%

HalidOdat · 2020-06-14T17:36:57Z

We still don't have benchmarks for the specialized objects 😅 , Like Number object not primitive, the benchmarks for string don't test String object only string primitives. although the speed boost is really nice :)

Edit: Made some benchmarks for some specialized objects #494

github-actions · 2020-06-14T20:08:22Z

Benchmark for `79188c1`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Array access (Execution)	25.8±0.79µs	37.0±1.89µs	-30.27%
Array creation (Execution)	7.0±0.14ms	15.2±0.22ms	-53.95%
Array pop (Execution)	2.5±0.05ms	9.2±0.24ms	-72.83%
Create Realm	174.5±4.96µs	503.1±21.23µs	-65.32%
Dynamic Object Property Access (Execution)	12.1±0.39µs	17.3±0.72µs	-30.06%
Expression (Lexer)	2.1±0.04µs	2.1±0.07µs	0.00%
Expression (Parser)	5.1±0.19µs	5.0±0.11µs	+2.00%
Fibonacci (Execution)	2.1±0.05ms	2.2±0.09ms	-4.55%
For loop (Execution)	58.4±2.59µs	66.2±3.50µs	-11.78%
For loop (Lexer)	5.3±0.07µs	5.4±0.16µs	-1.85%
For loop (Parser)	13.9±0.34µs	13.6±0.35µs	+2.21%
Hello World (Lexer)	930.5±17.10ns	933.0±38.21ns	-0.27%
Hello World (Parser)	2.2±0.06µs	2.3±0.07µs	-4.35%
Long file (Parser)	6.2±0.14ms	6.2±0.08ms	0.00%
Object Creation (Execution)	10.6±0.20µs	15.7±0.99µs	-32.48%
RegExp (Execution)	80.6±4.23µs	86.3±3.13µs	-6.60%
RegExp Literal (Execution)	81.9±1.54µs	94.8±5.87µs	-13.61%
RegExp Literal Creation (Execution)	77.4±1.94µs	88.0±5.57µs	-12.05%
Static Object Property Access (Execution)	11.0±0.41µs	16.5±0.86µs	-33.33%
String comparison (Execution)	11.1±0.38µs	20.3±1.17µs	-45.32%
String concatenation (Execution)	9.2±0.26µs	16.0±0.84µs	-42.50%
String copy (Execution)	7.7±0.18µs	13.4±0.76µs	-42.54%
Symbols (Execution)	8.7±0.28µs	17.9±0.61µs	-51.40%

jasonwilliams · 2020-06-15T08:18:55Z

We should do a before and after with the profiler, it would be nice to see

HalidOdat · 2020-06-15T17:58:57Z

Here is for master branch: https://gist.github.com/HalidOdat/c346ea541830a80f0042a716c6d9ecea

And for feature/specialized-objects branch: https://gist.github.com/HalidOdat/c46f9b6ca1c0e1a00aa29b514b4c3a69

github-actions · 2020-06-15T18:07:35Z

Benchmark for `3b70830`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Array access (Execution)	26.4±1.79µs	43.2±3.13µs	-38.89%
Array creation (Execution)	6.2±0.41ms	14.4±1.47ms	-56.94%
Array pop (Execution)	2.2±0.14ms	7.9±0.29ms	-72.15%
Create Realm	175.2±12.25µs	453.5±29.14µs	-61.37%
Dynamic Object Property Access (Execution)	11.8±0.64µs	16.8±1.15µs	-29.76%
Expression (Lexer)	2.0±0.14µs	1886.1±130.45ns	+6.04%
Expression (Parser)	4.4±0.35µs	4.4±0.30µs	0.00%
Fibonacci (Execution)	2.0±0.11ms	2.5±0.14ms	-20.00%
For loop (Execution)	56.8±3.44µs	67.2±6.31µs	-15.48%
For loop (Lexer)	4.8±0.31µs	4.9±0.34µs	-2.04%
For loop (Parser)	13.2±0.84µs	12.6±0.90µs	+4.76%
Hello World (Lexer)	901.6±64.70ns	887.5±75.48ns	+1.59%
Hello World (Parser)	2.0±0.14µs	2.1±0.14µs	-4.76%
Long file (Parser)	5.7±0.25ms	5.8±0.26ms	-1.72%
Object Creation (Execution)	10.6±0.68µs	15.3±1.95µs	-30.72%
RegExp (Execution)	77.2±6.14µs	80.7±4.90µs	-4.34%
RegExp Literal (Execution)	81.5±7.13µs	88.3±6.26µs	-7.70%
RegExp Literal Creation (Execution)	76.5±5.79µs	85.6±6.86µs	-10.63%
Static Object Property Access (Execution)	10.7±0.63µs	15.1±1.36µs	-29.14%
String comparison (Execution)	10.8±0.88µs	19.0±1.27µs	-43.16%
String concatenation (Execution)	9.0±0.69µs	14.7±0.86µs	-38.78%
String copy (Execution)	7.7±0.52µs	12.7±0.71µs	-39.37%
Symbols (Execution)	8.3±0.60µs	16.6±1.14µs	-50.00%

github-actions · 2020-06-15T18:09:16Z

Benchmark for `c9dcfd8`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Array access (Execution)	29.9±2.92µs	42.9±5.09µs	-30.30%
Array creation (Execution)	7.6±0.45ms	17.8±1.15ms	-57.30%
Array pop (Execution)	2.9±0.22ms	10.0±0.67ms	-71.00%
Create Realm	201.9±18.38µs	578.4±60.06µs	-65.09%
Dynamic Object Property Access (Execution)	14.5±1.58µs	21.2±2.55µs	-31.60%
Expression (Lexer)	2.4±0.22µs	2.6±0.34µs	-7.69%
Expression (Parser)	5.6±0.43µs	5.9±0.65µs	-5.08%
Fibonacci (Execution)	2.5±0.20ms	2.7±0.25ms	-7.41%
For loop (Execution)	73.6±8.94µs	75.3±5.57µs	-2.26%
For loop (Lexer)	5.9±0.64µs	6.4±0.82µs	-7.81%
For loop (Parser)	14.9±1.02µs	16.4±2.32µs	-9.15%
Hello World (Lexer)	1097.1±125.49ns	1117.3±131.36ns	-1.81%
Hello World (Parser)	2.5±0.18µs	2.6±0.29µs	-3.85%
Long file (Parser)	6.9±0.37ms	7.5±0.44ms	-8.00%
Object Creation (Execution)	13.5±1.66µs	18.9±2.41µs	-28.57%
RegExp (Execution)	97.2±14.62µs	113.0±16.63µs	-13.98%
RegExp Literal (Execution)	103.2±33.48µs	124.1±26.69µs	-16.84%
RegExp Literal Creation (Execution)	94.9±13.06µs	108.8±13.07µs	-12.78%
Static Object Property Access (Execution)	14.0±1.95µs	20.3±2.94µs	-31.03%
String comparison (Execution)	12.7±1.02µs	25.3±5.06µs	-49.80%
String concatenation (Execution)	10.9±1.04µs	19.0±2.80µs	-42.63%
String copy (Execution)	9.3±1.05µs	16.8±3.36µs	-44.64%
Symbols (Execution)	10.6±1.05µs	21.8±2.36µs	-51.38%

Razican

This is good to go from my side. We might want to wait for the extra benchmarks though, to have some extra info.

- Added helper functions for Object and ValueData - Added `Hash` trait to `Value` - Abstracted Object field access - Deprecated `ObjectInternalMethods` - Optimized Realm creation - Fixed `Symbol` primitive

hash The problem with using the pointer as the hash, it can become a security vulnerability.

This type change removes 8-bytes from Symbol type size.

- Added `#[inline]` to some small functions

github-actions · 2020-06-15T22:53:09Z

Benchmark for `79c718a`

Click to view benchmark

Test	PR Benchmark	Master Benchmark	%
Array access (Execution)	24.7±1.52µs	37.8±0.82µs	-34.66%
Array creation (Execution)	6.4±0.40ms	15.0±0.14ms	-57.33%
Array pop (Execution)	2.4±0.05ms	9.0±0.10ms	-73.33%
Boolean Object Access (Execution)	8.7±0.52µs	33.5±1.18µs	-74.03%
Create Realm	163.5±10.21µs	492.9±12.97µs	-66.83%
Dynamic Object Property Access (Execution)	11.1±0.71µs	17.3±0.45µs	-35.84%
Expression (Lexer)	2.1±0.03µs	2.1±0.04µs	0.00%
Expression (Parser)	4.8±0.28µs	5.0±0.07µs	-4.00%
Fibonacci (Execution)	1855.6±90.16µs	2.1±0.03ms	-11.64%
For loop (Execution)	52.8±3.32µs	64.4±1.75µs	-18.01%
For loop (Lexer)	4.8±0.26µs	5.3±0.13µs	-9.43%
For loop (Parser)	13.7±0.46µs	13.5±0.26µs	+1.48%
Hello World (Lexer)	855.4±50.32ns	922.0±34.64ns	-7.22%
Hello World (Parser)	2.1±0.14µs	2.2±0.04µs	-4.55%
Long file (Parser)	6.2±0.09ms	6.1±0.07ms	+1.64%
Number Object Access (Execution)	7.2±0.49µs	28.9±0.47µs	-75.09%
Object Creation (Execution)	9.7±0.52µs	15.6±0.52µs	-37.82%
RegExp (Execution)	70.4±3.14µs	85.2±2.24µs	-17.37%
RegExp Literal (Execution)	78.3±5.50µs	93.1±3.78µs	-15.90%
RegExp Literal Creation (Execution)	74.2±5.83µs	84.8±1.93µs	-12.50%
Static Object Property Access (Execution)	10.0±0.60µs	16.3±0.38µs	-38.65%
String Object Access (Execution)	15.2±1.14µs	47.5±0.79µs	-68.00%
String comparison (Execution)	11.1±0.31µs	20.2±0.49µs	-45.05%
String concatenation (Execution)	9.2±0.13µs	15.9±0.24µs	-42.14%
String copy (Execution)	7.7±0.22µs	13.2±0.28µs	-41.67%
Symbols (Execution)	8.0±0.49µs	18.4±0.47µs	-56.52%

HalidOdat · 2020-06-15T22:56:41Z

Lets merge this :)

HalidOdat added enhancement New feature or request performance Performance related changes and issues technical debt builtins PRs and Issues related to builtins/intrinsics labels May 25, 2020

HalidOdat changed the title ~~Initial object specialization of Number~~ Object specialization May 25, 2020

HalidOdat mentioned this pull request May 26, 2020

Fix function object constructable/callable #409

Merged

HalidOdat force-pushed the feature/specialized-objects branch from 928ee40 to 62bdd68 Compare May 26, 2020 11:37

HalidOdat force-pushed the feature/specialized-objects branch from 62bdd68 to 5c4dc6f Compare May 26, 2020 17:04

Razican reviewed May 27, 2020

View reviewed changes

boa/src/builtins/value/mod.rs Outdated Show resolved Hide resolved

hello2dj mentioned this pull request May 28, 2020

implement Map() #401

Closed

brian-gavin mentioned this pull request May 28, 2020

Implement Set() #400

Closed

HalidOdat added this to the v0.9.0 milestone May 28, 2020

HalidOdat force-pushed the feature/specialized-objects branch from 896ba46 to d3d7fb8 Compare May 30, 2020 12:52

This was referenced May 31, 2020

[Feature Request] - Create benchmark for Object creation and access #427

Closed

[Feature Request] - Create benchmark for Array manipulation #428

Closed

[Feature Request] - Create benchmark for RegExp manipulation #429

Closed

HalidOdat mentioned this pull request May 31, 2020

Profiler using measureme #317

Merged

HalidOdat requested review from Razican and jasonwilliams June 14, 2020 04:33

Razican reviewed Jun 14, 2020

View reviewed changes

HalidOdat force-pushed the feature/specialized-objects branch from d19ed61 to 598adfe Compare June 14, 2020 16:17

HalidOdat requested a review from Razican June 14, 2020 18:46

HalidOdat mentioned this pull request Jun 15, 2020

Refactoring Property flags #497

Closed

Razican approved these changes Jun 15, 2020

View reviewed changes

HalidOdat added 9 commits June 15, 2020 23:58

Specialized Number, String, Boolean, BigInt, Function, Symbol objects

81b7905

- Added helper functions for Object and ValueData - Added `Hash` trait to `Value` - Abstracted Object field access - Deprecated `ObjectInternalMethods` - Optimized Realm creation - Fixed `Symbol` primitive

Made Symbol unique hash be an incremented value instead of pointer

16ca21a

hash The problem with using the pointer as the hash, it can become a security vulnerability.

Changed Symbol description type from String to Box<str>

6c82241

This type change removes 8-bytes from Symbol type size.

Changed init functions to return a (&str, Value)

5eb1a39

Made Number constructor and methods more specification compliant.

d165c6c

Abstracted Math and Json objects

e03d349

Abstracted Object name and function object length.

71f2389

Made make_builtin_fn and make_constructor_fn take a usize

c144cbd

- Added `#[inline]` to some small functions

construct_type_error and construct_range_error

25572f5

HalidOdat force-pushed the feature/specialized-objects branch from ecd65b2 to 25572f5 Compare June 15, 2020 22:28

HalidOdat merged commit df13272 into master Jun 15, 2020

HalidOdat deleted the feature/specialized-objects branch June 15, 2020 22:56

HalidOdat mentioned this pull request Jun 18, 2020

Added undefined property to global scope #501

Merged

HalidOdat mentioned this pull request Jul 16, 2020

Fixed stack overflow bug when calling undefined field #568

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Object specialization #419

Object specialization #419

HalidOdat commented May 25, 2020 •

edited

Loading

github-actions bot commented May 26, 2020

github-actions bot commented May 26, 2020

HalidOdat commented May 26, 2020 •

edited

Loading

github-actions bot commented May 26, 2020

Razican commented May 26, 2020

github-actions bot commented May 27, 2020

HalidOdat commented May 27, 2020 •

edited

Loading

Razican commented May 27, 2020

jasonwilliams commented May 28, 2020

github-actions bot commented May 28, 2020

github-actions bot commented May 28, 2020

github-actions bot commented May 29, 2020

github-actions bot commented May 30, 2020

Razican left a comment

Razican Jun 14, 2020

Razican Jun 14, 2020

Razican Jun 14, 2020

github-actions bot commented Jun 14, 2020

HalidOdat commented Jun 14, 2020 •

edited

Loading

github-actions bot commented Jun 14, 2020

jasonwilliams commented Jun 15, 2020 •

edited

Loading

HalidOdat commented Jun 15, 2020 •

edited

Loading

github-actions bot commented Jun 15, 2020

github-actions bot commented Jun 15, 2020

Razican left a comment

github-actions bot commented Jun 15, 2020

HalidOdat commented Jun 15, 2020

		self.throw_type_error("cannot convert null or undefined to Object")?;
		unreachable!();

Object specialization #419

Object specialization #419

Conversation

HalidOdat commented May 25, 2020 • edited Loading

github-actions bot commented May 26, 2020

Benchmark for 3d93be3

github-actions bot commented May 26, 2020

Benchmark for 36cc48e

HalidOdat commented May 26, 2020 • edited Loading

github-actions bot commented May 26, 2020

Benchmark for b52be12

Razican commented May 26, 2020

github-actions bot commented May 27, 2020

Benchmark for b1ff14c

HalidOdat commented May 27, 2020 • edited Loading

Razican commented May 27, 2020

jasonwilliams commented May 28, 2020

github-actions bot commented May 28, 2020

Benchmark for d0b4745

github-actions bot commented May 28, 2020

Benchmark for 345bcc1

github-actions bot commented May 29, 2020

Benchmark for 38d0eb1

github-actions bot commented May 30, 2020

Benchmark for 494d109

Razican left a comment

Choose a reason for hiding this comment

Razican Jun 14, 2020

Choose a reason for hiding this comment

Razican Jun 14, 2020

Choose a reason for hiding this comment

Razican Jun 14, 2020

Choose a reason for hiding this comment

github-actions bot commented Jun 14, 2020

Benchmark for 31c4c6c

HalidOdat commented Jun 14, 2020 • edited Loading

github-actions bot commented Jun 14, 2020

Benchmark for 79188c1

jasonwilliams commented Jun 15, 2020 • edited Loading

HalidOdat commented Jun 15, 2020 • edited Loading

github-actions bot commented Jun 15, 2020

Benchmark for 3b70830

github-actions bot commented Jun 15, 2020

Benchmark for c9dcfd8

Razican left a comment

Choose a reason for hiding this comment

github-actions bot commented Jun 15, 2020

Benchmark for 79c718a

HalidOdat commented Jun 15, 2020

HalidOdat commented May 25, 2020 •

edited

Loading

Benchmark for `3d93be3`

Benchmark for `36cc48e`

HalidOdat commented May 26, 2020 •

edited

Loading

Benchmark for `b52be12`

Benchmark for `b1ff14c`

HalidOdat commented May 27, 2020 •

edited

Loading

Benchmark for `d0b4745`

Benchmark for `345bcc1`

Benchmark for `38d0eb1`

Benchmark for `494d109`

Benchmark for `31c4c6c`

HalidOdat commented Jun 14, 2020 •

edited

Loading

Benchmark for `79188c1`

jasonwilliams commented Jun 15, 2020 •

edited

Loading

HalidOdat commented Jun 15, 2020 •

edited

Loading

Benchmark for `3b70830`

Benchmark for `c9dcfd8`

Benchmark for `79c718a`