ladybird/Userland/Libraries/LibJS/Runtime/AbstractOperations.cpp

1016 lines
45 KiB
C++
Raw Normal View History

/*
* Copyright (c) 2020-2021, Linus Groh <linusg@serenityos.org>
* Copyright (c) 2021, Andreas Kling <kling@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#include <AK/CharacterTypes.h>
#include <AK/Function.h>
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
#include <AK/Optional.h>
#include <AK/TemporaryChange.h>
#include <AK/Utf16View.h>
#include <LibJS/Bytecode/Interpreter.h>
#include <LibJS/Interpreter.h>
#include <LibJS/Parser.h>
#include <LibJS/Runtime/AbstractOperations.h>
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
#include <LibJS/Runtime/Accessor.h>
#include <LibJS/Runtime/ArgumentsObject.h>
#include <LibJS/Runtime/Array.h>
#include <LibJS/Runtime/BoundFunction.h>
#include <LibJS/Runtime/Completion.h>
#include <LibJS/Runtime/DeclarativeEnvironment.h>
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
#include <LibJS/Runtime/ECMAScriptFunctionObject.h>
#include <LibJS/Runtime/ErrorTypes.h>
#include <LibJS/Runtime/FunctionEnvironment.h>
#include <LibJS/Runtime/FunctionObject.h>
#include <LibJS/Runtime/GlobalObject.h>
#include <LibJS/Runtime/Object.h>
#include <LibJS/Runtime/ObjectEnvironment.h>
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
#include <LibJS/Runtime/PropertyDescriptor.h>
#include <LibJS/Runtime/PropertyKey.h>
#include <LibJS/Runtime/ProxyObject.h>
#include <LibJS/Runtime/Reference.h>
namespace JS {
// 7.2.1 RequireObjectCoercible ( argument ), https://tc39.es/ecma262/#sec-requireobjectcoercible
ThrowCompletionOr<Value> require_object_coercible(GlobalObject& global_object, Value value)
{
auto& vm = global_object.vm();
if (value.is_nullish())
return vm.throw_completion<TypeError>(global_object, ErrorType::NotObjectCoercible, value.to_string_without_side_effects());
return value;
}
// 7.3.14 Call ( F, V [ , argumentsList ] ), https://tc39.es/ecma262/#sec-call
LibJS: Implement [[Call]] and [[Construct]] internal slots properly This patch implements: - Spec compliant [[Call]] and [[Construct]] internal slots, as virtual FunctionObject::internal_{call,construct}(). These effectively replace the old virtual FunctionObject::{call,construct}(), but with several advantages: - Clear and consistent naming, following the object internal methods - Use of completions - internal_construct() returns an Object, and not Value! This has been a source of confusion for a long time, since in the spec there's always an Object returned but the Value return type in LibJS meant that this could not be fully trusted and something could screw you over. - Arguments are passed explicitly in form of a MarkedValueList, allowing manipulation (BoundFunction). We still put them on the execution context as a lot of code depends on it (VM::arguments()), but not from the Call() / Construct() AOs anymore, which now allows for bypassing them and invoking [[Call]] / [[Construct]] directly. Nothing but Call() / Construct() themselves do that at the moment, but future additions to ECMA262 or already existing web specs might. - Spec compliant, standalone Call() and Construct() AOs: currently the closest we have is VM::{call,construct}(), but those try to cater to all the different function object subclasses at once, resulting in a horrible mess and calling AOs with functions they should never be called with; most prominently PrepareForOrdinaryCall and OrdinaryCallBindThis, which are only for ECMAScriptFunctionObject. As a result this also contains an implicit optimization: we no longer need to create a new function environment for NativeFunctions - which, worth mentioning, is what started this whole crusade in the first place :^)
2021-10-08 20:37:21 +01:00
ThrowCompletionOr<Value> call_impl(GlobalObject& global_object, Value function, Value this_value, Optional<MarkedValueList> arguments_list)
{
auto& vm = global_object.vm();
// 1. If argumentsList is not present, set argumentsList to a new empty List.
if (!arguments_list.has_value())
arguments_list = MarkedValueList { global_object.heap() };
// 2. If IsCallable(F) is false, throw a TypeError exception.
if (!function.is_function())
return vm.throw_completion<TypeError>(global_object, ErrorType::NotAFunction, function.to_string_without_side_effects());
// 3. Return ? F.[[Call]](V, argumentsList).
return function.as_function().internal_call(this_value, move(*arguments_list));
}
// 7.3.15 Construct ( F [ , argumentsList [ , newTarget ] ] ), https://tc39.es/ecma262/#sec-construct
LibJS: Implement [[Call]] and [[Construct]] internal slots properly This patch implements: - Spec compliant [[Call]] and [[Construct]] internal slots, as virtual FunctionObject::internal_{call,construct}(). These effectively replace the old virtual FunctionObject::{call,construct}(), but with several advantages: - Clear and consistent naming, following the object internal methods - Use of completions - internal_construct() returns an Object, and not Value! This has been a source of confusion for a long time, since in the spec there's always an Object returned but the Value return type in LibJS meant that this could not be fully trusted and something could screw you over. - Arguments are passed explicitly in form of a MarkedValueList, allowing manipulation (BoundFunction). We still put them on the execution context as a lot of code depends on it (VM::arguments()), but not from the Call() / Construct() AOs anymore, which now allows for bypassing them and invoking [[Call]] / [[Construct]] directly. Nothing but Call() / Construct() themselves do that at the moment, but future additions to ECMA262 or already existing web specs might. - Spec compliant, standalone Call() and Construct() AOs: currently the closest we have is VM::{call,construct}(), but those try to cater to all the different function object subclasses at once, resulting in a horrible mess and calling AOs with functions they should never be called with; most prominently PrepareForOrdinaryCall and OrdinaryCallBindThis, which are only for ECMAScriptFunctionObject. As a result this also contains an implicit optimization: we no longer need to create a new function environment for NativeFunctions - which, worth mentioning, is what started this whole crusade in the first place :^)
2021-10-08 20:37:21 +01:00
ThrowCompletionOr<Object*> construct(GlobalObject& global_object, FunctionObject& function, Optional<MarkedValueList> arguments_list, FunctionObject* new_target)
{
// 1. If newTarget is not present, set newTarget to F.
if (!new_target)
new_target = &function;
// 2. If argumentsList is not present, set argumentsList to a new empty List.
if (!arguments_list.has_value())
arguments_list = MarkedValueList { global_object.heap() };
// 3. Return ? F.[[Construct]](argumentsList, newTarget).
return function.internal_construct(move(*arguments_list), *new_target);
}
// 7.3.19 LengthOfArrayLike ( obj ), https://tc39.es/ecma262/#sec-lengthofarraylike
ThrowCompletionOr<size_t> length_of_array_like(GlobalObject& global_object, Object const& object)
{
auto& vm = global_object.vm();
auto result = TRY(object.get(vm.names.length));
return result.to_length(global_object);
}
// 7.3.20 CreateListFromArrayLike ( obj [ , elementTypes ] ), https://tc39.es/ecma262/#sec-createlistfromarraylike
ThrowCompletionOr<MarkedValueList> create_list_from_array_like(GlobalObject& global_object, Value value, Function<ThrowCompletionOr<void>(Value)> check_value)
{
auto& vm = global_object.vm();
auto& heap = global_object.heap();
// 1. If elementTypes is not present, set elementTypes to « Undefined, Null, Boolean, String, Symbol, Number, BigInt, Object ».
// 2. If Type(obj) is not Object, throw a TypeError exception.
if (!value.is_object())
return vm.throw_completion<TypeError>(global_object, ErrorType::NotAnObject, value.to_string_without_side_effects());
auto& array_like = value.as_object();
// 3. Let len be ? LengthOfArrayLike(obj).
auto length = TRY(length_of_array_like(global_object, array_like));
// 4. Let list be a new empty List.
auto list = MarkedValueList { heap };
// 5. Let index be 0.
// 6. Repeat, while index < len,
for (size_t i = 0; i < length; ++i) {
// a. Let indexName be ! ToString(𝔽(index)).
auto index_name = PropertyKey { i };
// b. Let next be ? Get(obj, indexName).
auto next = TRY(array_like.get(index_name));
// c. If Type(next) is not an element of elementTypes, throw a TypeError exception.
if (check_value)
TRY(check_value(next));
// d. Append next as the last element of list.
list.append(next);
}
// 7. Return list.
return ThrowCompletionOr(move(list));
}
// 7.3.23 SpeciesConstructor ( O, defaultConstructor ), https://tc39.es/ecma262/#sec-speciesconstructor
ThrowCompletionOr<FunctionObject*> species_constructor(GlobalObject& global_object, Object const& object, FunctionObject& default_constructor)
{
auto& vm = global_object.vm();
// 1. Let C be ? Get(O, "constructor").
auto constructor = TRY(object.get(vm.names.constructor));
// 2. If C is undefined, return defaultConstructor.
if (constructor.is_undefined())
return &default_constructor;
// 3. If Type(C) is not Object, throw a TypeError exception.
if (!constructor.is_object())
return vm.throw_completion<TypeError>(global_object, ErrorType::NotAConstructor, constructor.to_string_without_side_effects());
// 4. Let S be ? Get(C, @@species).
auto species = TRY(constructor.as_object().get(*vm.well_known_symbol_species()));
// 5. If S is either undefined or null, return defaultConstructor.
if (species.is_nullish())
return &default_constructor;
// 6. If IsConstructor(S) is true, return S.
if (species.is_constructor())
return &species.as_function();
// 7. Throw a TypeError exception.
return vm.throw_completion<TypeError>(global_object, ErrorType::NotAConstructor, species.to_string_without_side_effects());
}
// 7.3.25 GetFunctionRealm ( obj ), https://tc39.es/ecma262/#sec-getfunctionrealm
ThrowCompletionOr<Realm*> get_function_realm(GlobalObject& global_object, FunctionObject const& function)
{
auto& vm = global_object.vm();
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 1. Assert: ! IsCallable(obj) is true.
// 2. If obj has a [[Realm]] internal slot, then
if (function.realm()) {
// a. Return obj.[[Realm]].
return function.realm();
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
}
// 3. If obj is a bound function exotic object, then
if (is<BoundFunction>(function)) {
auto& bound_function = static_cast<BoundFunction const&>(function);
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// a. Let target be obj.[[BoundTargetFunction]].
auto& target = bound_function.bound_target_function();
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// b. Return ? GetFunctionRealm(target).
return get_function_realm(global_object, target);
}
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 4. If obj is a Proxy exotic object, then
if (is<ProxyObject>(function)) {
auto& proxy = static_cast<ProxyObject const&>(function);
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// a. If obj.[[ProxyHandler]] is null, throw a TypeError exception.
if (proxy.is_revoked())
return vm.throw_completion<TypeError>(global_object, ErrorType::ProxyRevoked);
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// b. Let proxyTarget be obj.[[ProxyTarget]].
auto& proxy_target = proxy.target();
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// c. Return ? GetFunctionRealm(proxyTarget).
VERIFY(proxy_target.is_function());
return get_function_realm(global_object, static_cast<FunctionObject const&>(proxy_target));
}
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 5. Return the current Realm Record.
return vm.current_realm();
}
// 8.5.2.1 InitializeBoundName ( name, value, environment ), 8.5.2.1 InitializeBoundName ( name, value, environment )
ThrowCompletionOr<void> initialize_bound_name(GlobalObject& global_object, FlyString const& name, Value value, Environment* environment)
{
auto& vm = global_object.vm();
// 1. If environment is not undefined, then
if (environment) {
// a. Perform environment.InitializeBinding(name, value).
MUST(environment->initialize_binding(global_object, name, value));
// b. Return NormalCompletion(undefined).
return {};
}
// 2. Else,
else {
// a. Let lhs be ResolveBinding(name).
auto lhs = vm.resolve_binding(name);
// b. Return ? PutValue(lhs, value).
return TRY(lhs.put_value(global_object, value));
}
VERIFY_NOT_REACHED();
}
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 10.1.6.2 IsCompatiblePropertyDescriptor ( Extensible, Desc, Current ), https://tc39.es/ecma262/#sec-iscompatiblepropertydescriptor
bool is_compatible_property_descriptor(bool extensible, PropertyDescriptor const& descriptor, Optional<PropertyDescriptor> const& current)
{
// 1. Return ValidateAndApplyPropertyDescriptor(undefined, undefined, Extensible, Desc, Current).
return validate_and_apply_property_descriptor(nullptr, {}, extensible, descriptor, current);
}
// 10.1.6.3 ValidateAndApplyPropertyDescriptor ( O, P, extensible, Desc, current ), https://tc39.es/ecma262/#sec-validateandapplypropertydescriptor
bool validate_and_apply_property_descriptor(Object* object, PropertyKey const& property_name, bool extensible, PropertyDescriptor const& descriptor, Optional<PropertyDescriptor> const& current)
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
{
// 1. Assert: If O is not undefined, then IsPropertyKey(P) is true.
if (object)
VERIFY(property_name.is_valid());
// 2. If current is undefined, then
if (!current.has_value()) {
// a. If extensible is false, return false.
if (!extensible)
return false;
// b. Assert: extensible is true.
// c. If IsGenericDescriptor(Desc) is true or IsDataDescriptor(Desc) is true, then
if (descriptor.is_generic_descriptor() || descriptor.is_data_descriptor()) {
// i. If O is not undefined, create an own data property named P of object O whose [[Value]], [[Writable]],
// [[Enumerable]], and [[Configurable]] attribute values are described by Desc.
// If the value of an attribute field of Desc is absent, the attribute of the newly created property is set
// to its default value.
if (object) {
auto value = descriptor.value.value_or(js_undefined());
object->storage_set(property_name, { value, descriptor.attributes() });
}
}
// d. Else,
else {
// i. Assert: ! IsAccessorDescriptor(Desc) is true.
VERIFY(descriptor.is_accessor_descriptor());
// ii. If O is not undefined, create an own accessor property named P of object O whose [[Get]], [[Set]],
// [[Enumerable]], and [[Configurable]] attribute values are described by Desc.
// If the value of an attribute field of Desc is absent, the attribute of the newly created property is set
// to its default value.
if (object) {
auto accessor = Accessor::create(object->vm(), descriptor.get.value_or(nullptr), descriptor.set.value_or(nullptr));
object->storage_set(property_name, { accessor, descriptor.attributes() });
}
}
// e. Return true.
return true;
}
// 3. If every field in Desc is absent, return true.
if (descriptor.is_empty())
return true;
// 4. If current.[[Configurable]] is false, then
if (!*current->configurable) {
// a. If Desc.[[Configurable]] is present and its value is true, return false.
if (descriptor.configurable.has_value() && *descriptor.configurable)
return false;
// b. If Desc.[[Enumerable]] is present and ! SameValue(Desc.[[Enumerable]], current.[[Enumerable]]) is false, return false.
if (descriptor.enumerable.has_value() && *descriptor.enumerable != *current->enumerable)
return false;
}
// 5. If ! IsGenericDescriptor(Desc) is true, then
if (descriptor.is_generic_descriptor()) {
// a. NOTE: No further validation is required.
}
// 6. Else if ! SameValue(! IsDataDescriptor(current), ! IsDataDescriptor(Desc)) is false, then
else if (current->is_data_descriptor() != descriptor.is_data_descriptor()) {
// a. If current.[[Configurable]] is false, return false.
if (!*current->configurable)
return false;
// b. If IsDataDescriptor(current) is true, then
if (current->is_data_descriptor()) {
// If O is not undefined, convert the property named P of object O from a data property to an accessor property.
// Preserve the existing values of the converted property's [[Configurable]] and [[Enumerable]] attributes and
// set the rest of the property's attributes to their default values.
if (object) {
auto accessor = Accessor::create(object->vm(), nullptr, nullptr);
object->storage_set(property_name, { accessor, current->attributes() });
}
}
// c. Else,
else {
// If O is not undefined, convert the property named P of object O from an accessor property to a data property.
// Preserve the existing values of the converted property's [[Configurable]] and [[Enumerable]] attributes and
// set the rest of the property's attributes to their default values.
if (object) {
auto value = js_undefined();
object->storage_set(property_name, { value, current->attributes() });
}
}
}
// 7. Else if IsDataDescriptor(current) and IsDataDescriptor(Desc) are both true, then
else if (current->is_data_descriptor() && descriptor.is_data_descriptor()) {
// a. If current.[[Configurable]] is false and current.[[Writable]] is false, then
if (!*current->configurable && !*current->writable) {
// i. If Desc.[[Writable]] is present and Desc.[[Writable]] is true, return false.
if (descriptor.writable.has_value() && *descriptor.writable)
return false;
// ii. If Desc.[[Value]] is present and SameValue(Desc.[[Value]], current.[[Value]]) is false, return false.
if (descriptor.value.has_value() && !same_value(*descriptor.value, *current->value))
return false;
// iii. Return true.
return true;
}
}
// 8. Else,
else {
// a. Assert: ! IsAccessorDescriptor(current) and ! IsAccessorDescriptor(Desc) are both true.
VERIFY(current->is_accessor_descriptor());
VERIFY(descriptor.is_accessor_descriptor());
// b. If current.[[Configurable]] is false, then
if (!*current->configurable) {
// i. If Desc.[[Set]] is present and SameValue(Desc.[[Set]], current.[[Set]]) is false, return false.
if (descriptor.set.has_value() && *descriptor.set != *current->set)
return false;
// ii. If Desc.[[Get]] is present and SameValue(Desc.[[Get]], current.[[Get]]) is false, return false.
if (descriptor.get.has_value() && *descriptor.get != *current->get)
return false;
// iii. Return true.
return true;
}
}
// 9. If O is not undefined, then
if (object) {
// a. For each field of Desc that is present, set the corresponding attribute of the property named P of object O to the value of the field.
Value value;
if (descriptor.is_accessor_descriptor() || (current->is_accessor_descriptor() && !descriptor.is_data_descriptor())) {
auto* getter = descriptor.get.value_or(current->get.value_or(nullptr));
auto* setter = descriptor.set.value_or(current->set.value_or(nullptr));
value = Accessor::create(object->vm(), getter, setter);
} else {
value = descriptor.value.value_or(current->value.value_or({}));
}
PropertyAttributes attributes;
attributes.set_writable(descriptor.writable.value_or(current->writable.value_or(false)));
attributes.set_enumerable(descriptor.enumerable.value_or(current->enumerable.value_or(false)));
attributes.set_configurable(descriptor.configurable.value_or(current->configurable.value_or(false)));
object->storage_set(property_name, { value, attributes });
}
// 10. Return true.
return true;
}
2021-08-08 11:04:30 +01:00
// 10.1.14 GetPrototypeFromConstructor ( constructor, intrinsicDefaultProto ), https://tc39.es/ecma262/#sec-getprototypefromconstructor
ThrowCompletionOr<Object*> get_prototype_from_constructor(GlobalObject& global_object, FunctionObject const& constructor, Object* (GlobalObject::*intrinsic_default_prototype)())
{
auto& vm = global_object.vm();
// 1. Assert: intrinsicDefaultProto is this specification's name of an intrinsic object. The corresponding object must be an intrinsic that is intended to be used as the [[Prototype]] value of an object.
// 2. Let proto be ? Get(constructor, "prototype").
auto prototype = TRY(constructor.get(vm.names.prototype));
// 3. If Type(proto) is not Object, then
if (!prototype.is_object()) {
// a. Let realm be ? GetFunctionRealm(constructor).
auto* realm = TRY(get_function_realm(global_object, constructor));
// b. Set proto to realm's intrinsic object named intrinsicDefaultProto.
prototype = (realm->global_object().*intrinsic_default_prototype)();
}
// 4. Return proto.
return &prototype.as_object();
}
// 9.1.2.2 NewDeclarativeEnvironment ( E ), https://tc39.es/ecma262/#sec-newdeclarativeenvironment
DeclarativeEnvironment* new_declarative_environment(Environment& environment)
{
auto& global_object = environment.global_object();
return global_object.heap().allocate<DeclarativeEnvironment>(global_object, &environment);
}
// 9.1.2.3 NewObjectEnvironment ( O, W, E ), https://tc39.es/ecma262/#sec-newobjectenvironment
ObjectEnvironment* new_object_environment(Object& object, bool is_with_environment, Environment* environment)
{
auto& global_object = object.global_object();
return global_object.heap().allocate<ObjectEnvironment>(global_object, object, is_with_environment ? ObjectEnvironment::IsWithEnvironment::Yes : ObjectEnvironment::IsWithEnvironment::No, environment);
}
// 9.1.2.4 NewFunctionEnvironment ( F, newTarget ), https://tc39.es/ecma262/#sec-newfunctionenvironment
FunctionEnvironment* new_function_environment(ECMAScriptFunctionObject& function, Object* new_target)
{
auto& global_object = function.global_object();
// 1. Let env be a new function Environment Record containing no bindings.
auto* env = global_object.heap().allocate<FunctionEnvironment>(global_object, function.environment());
// 2. Set env.[[FunctionObject]] to F.
env->set_function_object(function);
// 3. If F.[[ThisMode]] is lexical, set env.[[ThisBindingStatus]] to lexical.
if (function.this_mode() == ECMAScriptFunctionObject::ThisMode::Lexical)
env->set_this_binding_status(FunctionEnvironment::ThisBindingStatus::Lexical);
// 4. Else, set env.[[ThisBindingStatus]] to uninitialized.
else
env->set_this_binding_status(FunctionEnvironment::ThisBindingStatus::Uninitialized);
// 5. Set env.[[NewTarget]] to newTarget.
env->set_new_target(new_target ?: js_undefined());
// 6. Set env.[[OuterEnv]] to F.[[Environment]].
// NOTE: Done in step 1 via the FunctionEnvironment constructor.
// 7. Return env.
return env;
}
2021-10-11 20:29:31 +02:00
PrivateEnvironment* new_private_environment(VM& vm, PrivateEnvironment* outer)
{
return vm.heap().allocate<PrivateEnvironment>(vm.current_realm()->global_object(), outer);
}
// 9.4.3 GetThisEnvironment ( ), https://tc39.es/ecma262/#sec-getthisenvironment
Environment& get_this_environment(VM& vm)
{
for (auto* env = vm.lexical_environment(); env; env = env->outer_environment()) {
if (env->has_this_binding())
return *env;
}
VERIFY_NOT_REACHED();
}
// 13.3.7.2 GetSuperConstructor ( ), https://tc39.es/ecma262/#sec-getsuperconstructor
Object* get_super_constructor(VM& vm)
{
// 1. Let envRec be GetThisEnvironment().
auto& env = get_this_environment(vm);
// 2. Assert: envRec is a function Environment Record.
// 3. Let activeFunction be envRec.[[FunctionObject]].
// 4. Assert: activeFunction is an ECMAScript function object.
auto& active_function = verify_cast<FunctionEnvironment>(env).function_object();
// 5. Let superConstructor be ! activeFunction.[[GetPrototypeOf]]().
2021-10-03 15:57:26 +01:00
auto* super_constructor = MUST(active_function.internal_get_prototype_of());
// 6. Return superConstructor.
return super_constructor;
}
2021-08-08 11:04:30 +01:00
// 13.3.7.3 MakeSuperPropertyReference ( actualThis, propertyKey, strict ), https://tc39.es/ecma262/#sec-makesuperpropertyreference
ThrowCompletionOr<Reference> make_super_property_reference(GlobalObject& global_object, Value actual_this, PropertyKey const& property_key, bool strict)
{
auto& vm = global_object.vm();
// 1. Let env be GetThisEnvironment().
auto& env = verify_cast<FunctionEnvironment>(get_this_environment(vm));
// 2. Assert: env.HasSuperBinding() is true.
VERIFY(env.has_super_binding());
// 3. Let baseValue be ? env.GetSuperBase().
auto base_value = TRY(env.get_super_base());
// 4. Let bv be ? RequireObjectCoercible(baseValue).
auto bv = TRY(require_object_coercible(global_object, base_value));
// 5. Return the Reference Record { [[Base]]: bv, [[ReferencedName]]: propertyKey, [[Strict]]: strict, [[ThisValue]]: actualThis }.
// 6. NOTE: This returns a Super Reference Record.
return Reference { bv, property_key, actual_this, strict };
}
// 19.2.1.1 PerformEval ( x, callerRealm, strictCaller, direct ), https://tc39.es/ecma262/#sec-performeval
ThrowCompletionOr<Value> perform_eval(Value x, GlobalObject& caller_realm, CallerMode strict_caller, EvalMode direct)
{
VERIFY(direct == EvalMode::Direct || strict_caller == CallerMode::NonStrict);
if (!x.is_string())
return x;
auto& vm = caller_realm.vm();
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
auto& eval_realm = vm.running_execution_context().realm;
auto& code_string = x.as_string();
Parser parser { Lexer { code_string.string() } };
auto program = parser.parse_program(strict_caller == CallerMode::Strict);
if (parser.has_errors()) {
auto& error = parser.errors()[0];
return vm.throw_completion<SyntaxError>(caller_realm, error.to_string());
}
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
auto strict_eval = strict_caller == CallerMode::Strict;
if (program->is_strict_mode())
strict_eval = true;
auto& running_context = vm.running_execution_context();
Environment* lexical_environment;
Environment* variable_environment;
2021-10-11 20:29:31 +02:00
PrivateEnvironment* private_environment;
if (direct == EvalMode::Direct) {
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
lexical_environment = new_declarative_environment(*running_context.lexical_environment);
variable_environment = running_context.variable_environment;
2021-10-11 20:29:31 +02:00
private_environment = running_context.private_environment;
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
} else {
lexical_environment = new_declarative_environment(eval_realm->global_environment());
variable_environment = &eval_realm->global_environment();
2021-10-11 20:29:31 +02:00
private_environment = nullptr;
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
}
if (strict_eval)
variable_environment = lexical_environment;
if (direct == EvalMode::Direct && !strict_eval) {
// NOTE: Non-strict direct eval() forces us to deoptimize variable accesses.
// Mark the variable environment chain as screwed since we will not be able
// to rely on cached environment coordinates from this point on.
variable_environment->set_permanently_screwed_by_eval();
}
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
// 18. If runningContext is not already suspended, suspend runningContext.
// FIXME: We don't have this concept yet.
ExecutionContext eval_context(vm.heap());
eval_context.realm = eval_realm;
eval_context.variable_environment = variable_environment;
eval_context.lexical_environment = lexical_environment;
2021-10-11 20:29:31 +02:00
eval_context.private_environment = private_environment;
TRY(vm.push_execution_context(eval_context, eval_realm->global_object()));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
ScopeGuard pop_guard = [&] {
vm.pop_execution_context();
};
2021-10-11 20:29:31 +02:00
TRY(eval_declaration_instantiation(vm, eval_realm->global_object(), program, variable_environment, lexical_environment, private_environment, strict_eval));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
TemporaryChange scope_change_strict(vm.running_execution_context().is_strict_mode, strict_eval);
Value eval_result;
if (auto* bytecode_interpreter = Bytecode::Interpreter::current()) {
auto executable = JS::Bytecode::Generator::generate(program);
executable.name = "eval"sv;
if (JS::Bytecode::g_dump_bytecode)
executable.dump();
eval_result = TRY(bytecode_interpreter->run(executable));
} else {
auto& ast_interpreter = vm.interpreter();
// FIXME: We need to use evaluate_statements() here because Program::execute() calls global_declaration_instantiation() when it shouldn't
eval_result = program->evaluate_statements(ast_interpreter, caller_realm);
}
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
if (auto* exception = vm.exception())
return throw_completion(exception->value());
else
return eval_result.value_or(js_undefined());
}
// 19.2.1.3 EvalDeclarationInstantiation ( body, varEnv, lexEnv, privateEnv, strict ), https://tc39.es/ecma262/#sec-evaldeclarationinstantiation
2021-10-11 20:29:31 +02:00
ThrowCompletionOr<void> eval_declaration_instantiation(VM& vm, GlobalObject& global_object, Program const& program, Environment* variable_environment, Environment* lexical_environment, PrivateEnvironment* private_environment, bool strict)
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
{
// FIXME: I'm not sure if the global object is correct here. And this is quite a crucial spot!
GlobalEnvironment* global_var_environment = variable_environment->is_global_environment() ? static_cast<GlobalEnvironment*>(variable_environment) : nullptr;
if (!strict) {
if (global_var_environment) {
program.for_each_var_declared_name([&](auto const& name) {
if (global_var_environment->has_lexical_declaration(name)) {
vm.throw_exception<SyntaxError>(global_object, ErrorType::TopLevelVariableAlreadyDeclared, name);
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
return IterationDecision::Break;
}
return IterationDecision::Continue;
});
}
auto* this_environment = lexical_environment;
while (this_environment != variable_environment) {
if (!is<ObjectEnvironment>(*this_environment)) {
program.for_each_var_declared_name([&](auto const& name) {
if (MUST(this_environment->has_binding(name))) {
vm.throw_exception<SyntaxError>(global_object, ErrorType::TopLevelVariableAlreadyDeclared, name);
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
return IterationDecision::Break;
}
// FIXME: NOTE: Annex B.3.4 defines alternate semantics for the above step.
// In particular it only throw the syntax error if it is not an environment from a catchclause.
return IterationDecision::Continue;
});
if (auto* exception = vm.exception())
return throw_completion(exception->value());
}
this_environment = this_environment->outer_environment();
VERIFY(this_environment);
}
}
2021-10-11 20:29:31 +02:00
// FIXME: Add Private identifiers check here.
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
HashTable<FlyString> declared_function_names;
Vector<FunctionDeclaration const&> functions_to_initialize;
program.for_each_var_function_declaration_in_reverse_order([&](FunctionDeclaration const& function) {
if (declared_function_names.set(function.name()) != AK::HashSetResult::InsertedNewEntry)
return IterationDecision::Continue;
if (global_var_environment) {
auto function_definable = global_var_environment->can_declare_global_function(function.name());
if (vm.exception())
return IterationDecision::Break;
if (!function_definable) {
vm.throw_exception<TypeError>(global_object, ErrorType::CannotDeclareGlobalFunction, function.name());
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
return IterationDecision::Break;
}
}
functions_to_initialize.append(function);
return IterationDecision::Continue;
});
if (auto* exception = vm.exception())
return throw_completion(exception->value());
if (!strict) {
// The spec here uses 'declaredVarNames' but that has not been declared yet.
HashTable<FlyString> hoisted_functions;
program.for_each_function_hoistable_with_annexB_extension([&](FunctionDeclaration& function_declaration) {
auto& function_name = function_declaration.name();
auto* this_environment = lexical_environment;
while (this_environment != variable_environment) {
if (!is<ObjectEnvironment>(*this_environment) && MUST(this_environment->has_binding(function_name)))
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
return IterationDecision::Continue;
this_environment = this_environment->outer_environment();
VERIFY(this_environment);
}
if (global_var_environment) {
if (global_var_environment->has_lexical_declaration(function_name))
return IterationDecision::Continue;
auto var_definable = global_var_environment->can_declare_global_var(function_name);
if (vm.exception())
return IterationDecision::Break;
if (!var_definable)
return IterationDecision::Continue;
}
if (!declared_function_names.contains(function_name) && !hoisted_functions.contains(function_name)) {
if (global_var_environment) {
global_var_environment->create_global_var_binding(function_name, true);
if (vm.exception())
return IterationDecision::Break;
} else {
if (!MUST(variable_environment->has_binding(function_name))) {
MUST(variable_environment->create_mutable_binding(global_object, function_name, true));
MUST(variable_environment->initialize_binding(global_object, function_name, js_undefined()));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
}
}
hoisted_functions.set(function_name);
}
function_declaration.set_should_do_additional_annexB_steps();
return IterationDecision::Continue;
});
if (auto* exception = vm.exception())
return throw_completion(exception->value());
}
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
HashTable<FlyString> declared_var_names;
program.for_each_var_scoped_variable_declaration([&](VariableDeclaration const& declaration) {
declaration.for_each_bound_name([&](auto const& name) {
if (!declared_function_names.contains(name)) {
if (global_var_environment) {
auto variable_definable = global_var_environment->can_declare_global_var(name);
if (vm.exception())
return IterationDecision::Break;
if (!variable_definable) {
vm.throw_exception<TypeError>(global_object, ErrorType::CannotDeclareGlobalVariable, name);
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
return IterationDecision::Break;
}
}
declared_var_names.set(name);
}
return IterationDecision::Continue;
});
if (vm.exception())
return IterationDecision::Break;
return IterationDecision::Continue;
});
if (auto* exception = vm.exception())
return throw_completion(exception->value());
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
// 14. NOTE: No abnormal terminations occur after this algorithm step unless varEnv is a global Environment Record and the global object is a Proxy exotic object.
program.for_each_lexically_scoped_declaration([&](Declaration const& declaration) {
declaration.for_each_bound_name([&](auto const& name) {
if (declaration.is_constant_declaration())
(void)lexical_environment->create_immutable_binding(global_object, name, true);
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
else
(void)lexical_environment->create_mutable_binding(global_object, name, false);
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
if (vm.exception())
return IterationDecision::Break;
return IterationDecision::Continue;
});
if (vm.exception())
return IterationDecision::Break;
return IterationDecision::Continue;
});
if (auto* exception = vm.exception())
return throw_completion(exception->value());
for (auto& declaration : functions_to_initialize) {
2021-10-11 20:29:31 +02:00
auto* function = ECMAScriptFunctionObject::create(global_object, declaration.name(), declaration.body(), declaration.parameters(), declaration.function_length(), lexical_environment, private_environment, declaration.kind(), declaration.is_strict_mode(), declaration.might_need_arguments_object());
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
if (global_var_environment) {
global_var_environment->create_global_function_binding(declaration.name(), function, true);
if (auto* exception = vm.exception())
return throw_completion(exception->value());
} else {
auto binding_exists = MUST(variable_environment->has_binding(declaration.name()));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
if (!binding_exists) {
TRY(variable_environment->create_mutable_binding(global_object, declaration.name(), true));
TRY(variable_environment->initialize_binding(global_object, declaration.name(), function));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
} else {
TRY(variable_environment->set_mutable_binding(global_object, declaration.name(), function, false));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
}
}
}
for (auto& var_name : declared_var_names) {
if (global_var_environment) {
global_var_environment->create_global_var_binding(var_name, true);
if (auto* exception = vm.exception())
return throw_completion(exception->value());
} else {
auto binding_exists = MUST(variable_environment->has_binding(var_name));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
if (!binding_exists) {
TRY(variable_environment->create_mutable_binding(global_object, var_name, true));
TRY(variable_environment->initialize_binding(global_object, var_name, js_undefined()));
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
}
}
}
return {};
}
// 10.4.4.6 CreateUnmappedArgumentsObject ( argumentsList ), https://tc39.es/ecma262/#sec-createunmappedargumentsobject
Object* create_unmapped_arguments_object(GlobalObject& global_object, Span<Value> arguments)
{
auto& vm = global_object.vm();
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 1. Let len be the number of elements in argumentsList.
auto length = arguments.size();
// 2. Let obj be ! OrdinaryObjectCreate(%Object.prototype%, « [[ParameterMap]] »).
// 3. Set obj.[[ParameterMap]] to undefined.
auto* object = Object::create(global_object, global_object.object_prototype());
object->set_has_parameter_map();
// 4. Perform DefinePropertyOrThrow(obj, "length", PropertyDescriptor { [[Value]]: 𝔽(len), [[Writable]]: true, [[Enumerable]]: false, [[Configurable]]: true }).
MUST(object->define_property_or_throw(vm.names.length, { .value = Value(length), .writable = true, .enumerable = false, .configurable = true }));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 5. Let index be 0.
// 6. Repeat, while index < len,
for (size_t index = 0; index < length; ++index) {
// a. Let val be argumentsList[index].
auto value = arguments[index];
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// b. Perform ! CreateDataPropertyOrThrow(obj, ! ToString(𝔽(index)), val).
MUST(object->create_data_property_or_throw(index, value));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// c. Set index to index + 1.
}
// 7. Perform ! DefinePropertyOrThrow(obj, @@iterator, PropertyDescriptor { [[Value]]: %Array.prototype.values%, [[Writable]]: true, [[Enumerable]]: false, [[Configurable]]: true }).
auto* array_prototype_values = global_object.array_prototype_values_function();
MUST(object->define_property_or_throw(*vm.well_known_symbol_iterator(), { .value = array_prototype_values, .writable = true, .enumerable = false, .configurable = true }));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 8. Perform ! DefinePropertyOrThrow(obj, "callee", PropertyDescriptor { [[Get]]: %ThrowTypeError%, [[Set]]: %ThrowTypeError%, [[Enumerable]]: false, [[Configurable]]: false }).
auto* throw_type_error = global_object.throw_type_error_function();
MUST(object->define_property_or_throw(vm.names.callee, { .get = throw_type_error, .set = throw_type_error, .enumerable = false, .configurable = false }));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 9. Return obj.
return object;
}
// 10.4.4.7 CreateMappedArgumentsObject ( func, formals, argumentsList, env ), https://tc39.es/ecma262/#sec-createmappedargumentsobject
Object* create_mapped_arguments_object(GlobalObject& global_object, FunctionObject& function, Vector<FunctionNode::Parameter> const& formals, Span<Value> arguments, Environment& environment)
{
auto& vm = global_object.vm();
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 1. Assert: formals does not contain a rest parameter, any binding patterns, or any initializers. It may contain duplicate identifiers.
// 2. Let len be the number of elements in argumentsList.
VERIFY(arguments.size() <= NumericLimits<i32>::max());
i32 length = static_cast<i32>(arguments.size());
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 3. Let obj be ! MakeBasicObject(« [[Prototype]], [[Extensible]], [[ParameterMap]] »).
// 4. Set obj.[[GetOwnProperty]] as specified in 10.4.4.1.
// 5. Set obj.[[DefineOwnProperty]] as specified in 10.4.4.2.
// 6. Set obj.[[Get]] as specified in 10.4.4.3.
// 7. Set obj.[[Set]] as specified in 10.4.4.4.
// 8. Set obj.[[Delete]] as specified in 10.4.4.5.
// 9. Set obj.[[Prototype]] to %Object.prototype%.
auto* object = vm.heap().allocate<ArgumentsObject>(global_object, global_object, environment);
VERIFY(!vm.exception());
// 14. Let index be 0.
// 15. Repeat, while index < len,
for (i32 index = 0; index < length; ++index) {
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// a. Let val be argumentsList[index].
auto value = arguments[index];
// b. Perform ! CreateDataPropertyOrThrow(obj, ! ToString(𝔽(index)), val).
MUST(object->create_data_property_or_throw(index, value));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// c. Set index to index + 1.
}
// 16. Perform ! DefinePropertyOrThrow(obj, "length", PropertyDescriptor { [[Value]]: 𝔽(len), [[Writable]]: true, [[Enumerable]]: false, [[Configurable]]: true }).
MUST(object->define_property_or_throw(vm.names.length, { .value = Value(length), .writable = true, .enumerable = false, .configurable = true }));
// 17. Let mappedNames be a new empty List.
HashTable<FlyString> mapped_names;
// 18. Set index to numberOfParameters - 1.
// 19. Repeat, while index ≥ 0,
VERIFY(formals.size() <= NumericLimits<i32>::max());
for (i32 index = static_cast<i32>(formals.size()) - 1; index >= 0; --index) {
// a. Let name be parameterNames[index].
auto const& name = formals[index].binding.get<FlyString>();
// b. If name is not an element of mappedNames, then
if (mapped_names.contains(name))
continue;
// i. Add name as an element of the list mappedNames.
mapped_names.set(name);
// ii. If index < len, then
if (index < length) {
// 1. Let g be MakeArgGetter(name, env).
// 2. Let p be MakeArgSetter(name, env).
// 3. Perform map.[[DefineOwnProperty]](! ToString(𝔽(index)), PropertyDescriptor { [[Set]]: p, [[Get]]: g, [[Enumerable]]: false, [[Configurable]]: true }).
object->parameter_map().define_native_accessor(
PropertyKey { index },
[&environment, name](VM&, GlobalObject& global_object_getter) -> JS::ThrowCompletionOr<Value> {
return MUST(environment.get_binding_value(global_object_getter, name, false));
},
LibJS: Make scoping follow the spec Before this we used an ad-hoc combination of references and 'variables' stored in a hashmap. This worked in most cases but is not spec like. Additionally hoisting, dynamically naming functions and scope analysis was not done properly. This patch fixes all of that by: - Implement BindingInitialization for destructuring assignment. - Implementing a new ScopePusher which tracks the lexical and var scoped declarations. This hoists functions to the top level if no lexical declaration name overlaps. Furthermore we do checking of redeclarations in the ScopePusher now requiring less checks all over the place. - Add methods for parsing the directives and statement lists instead of having that code duplicated in multiple places. This allows declarations to pushed to the appropriate scope more easily. - Remove the non spec way of storing 'variables' in DeclarativeEnvironment and make Reference follow the spec instead of checking both the bindings and 'variables'. - Remove all scoping related things from the Interpreter. And instead use environments as specified by the spec. This also includes fixing that NativeFunctions did not produce a valid FunctionEnvironment which could cause issues with callbacks and eval. All FunctionObjects now have a valid NewFunctionEnvironment implementation. - Remove execute_statements from Interpreter and instead use ASTNode::execute everywhere this simplifies AST.cpp as you no longer need to worry about which method to call. - Make ScopeNodes setup their own environment. This uses four different methods specified by the spec {Block, Function, Eval, Global}DeclarationInstantiation with the annexB extensions. - Implement and use NamedEvaluation where specified. Additionally there are fixes to things exposed by these changes to eval, {for, for-in, for-of} loops and assignment. Finally it also fixes some tests in test-js which where passing before but not now that we have correct behavior :^).
2021-09-22 12:44:56 +02:00
[&environment, name](VM& vm, GlobalObject& global_object_setter) {
MUST(environment.set_mutable_binding(global_object_setter, name, vm.argument(0), false));
return js_undefined();
},
Attribute::Configurable);
}
}
// 20. Perform ! DefinePropertyOrThrow(obj, @@iterator, PropertyDescriptor { [[Value]]: %Array.prototype.values%, [[Writable]]: true, [[Enumerable]]: false, [[Configurable]]: true }).
auto* array_prototype_values = global_object.array_prototype_values_function();
MUST(object->define_property_or_throw(*vm.well_known_symbol_iterator(), { .value = array_prototype_values, .writable = true, .enumerable = false, .configurable = true }));
// 21. Perform ! DefinePropertyOrThrow(obj, "callee", PropertyDescriptor { [[Value]]: func, [[Writable]]: true, [[Enumerable]]: false, [[Configurable]]: true }).
MUST(object->define_property_or_throw(vm.names.callee, { .value = &function, .writable = true, .enumerable = false, .configurable = true }));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 22. Return obj.
return object;
}
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 7.1.21 CanonicalNumericIndexString ( argument ), https://tc39.es/ecma262/#sec-canonicalnumericindexstring
Value canonical_numeric_index_string(GlobalObject& global_object, PropertyKey const& property_name)
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
{
// NOTE: If the property name is a number type (An implementation-defined optimized
// property key type), it can be treated as a string property that has already been
// converted successfully into a canonical numeric index.
VERIFY(property_name.is_string() || property_name.is_number());
if (property_name.is_number())
return Value(property_name.as_number());
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 1. Assert: Type(argument) is String.
auto argument = Value(js_string(global_object.vm(), property_name.as_string()));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 2. If argument is "-0", return -0𝔽.
if (argument.as_string().string() == "-0")
return Value(-0.0);
// 3. Let n be ! ToNumber(argument).
auto n = MUST(argument.to_number(global_object));
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
// 4. If SameValue(! ToString(n), argument) is false, return undefined.
if (!same_value(MUST(n.to_primitive_string(global_object)), argument))
LibJS: Rewrite most of Object for spec compliance :^) This is a huge patch, I know. In hindsight this perhaps could've been done slightly more incremental, but I started and then fixed everything until it worked, and here we are. I tried splitting of some completely unrelated changes into separate commits, however. Anyway. This is a rewrite of most of Object, and by extension large parts of Array, Proxy, Reflect, String, TypedArray, and some other things. What we already had worked fine for about 90% of things, but getting the last 10% right proved to be increasingly difficult with the current code that sort of grew organically and is only very loosely based on the spec - this became especially obvious when we started fixing a large number of test262 failures. Key changes include: - 1:1 matching function names and parameters of all object-related functions, to avoid ambiguity. Previously we had things like put(), which the spec doesn't have - as a result it wasn't always clear which need to be used. - Better separation between object abstract operations and internal methods - the former are always the same, the latter can be overridden (and are therefore virtual). The internal methods (i.e. [[Foo]] in the spec) are now prefixed with 'internal_' for clarity - again, it was previously not always clear which AO a certain method represents, get() could've been both Get and [[Get]] (I don't know which one it was closer to right now). Note that some of the old names have been kept until all code relying on them is updated, but they are now simple wrappers around the closest matching standard abstract operation. - Simplifications of the storage layer: functions that write values to storage are now prefixed with 'storage_' to make their purpose clear, and as they are not part of the spec they should not contain any steps specified by it. Much functionality is now covered by the layers above it and was removed (e.g. handling of accessors, attribute checks). - PropertyAttributes has been greatly simplified, and is being replaced by PropertyDescriptor - a concept similar to the current implementation, but more aligned with the actual spec. See the commit message of the previous commit where it was introduced for details. - As a bonus, and since I had to look at the spec a whole lot anyway, I introduced more inline comments with the exact steps from the spec - this makes it super easy to verify correctness. - East-const all the things. As a result of all of this, things are much more correct but a bit slower now. Retaining speed wasn't a consideration at all, I have done no profiling of the new code - there might be low hanging fruits, which we can then harvest separately. Special thanks to Idan for helping me with this by tracking down bugs, updating everything outside of LibJS to work with these changes (LibWeb, Spreadsheet, HackStudio), as well as providing countless patches to fix regressions I introduced - there still are very few (we got it down to 5), but we also get many new passing test262 tests in return. :^) Co-authored-by: Idan Horowitz <idan.horowitz@gmail.com>
2021-07-04 18:14:16 +01:00
return js_undefined();
// 5. Return n.
return n;
}
// 22.1.3.17.1 GetSubstitution ( matched, str, position, captures, namedCaptures, replacement ), https://tc39.es/ecma262/#sec-getsubstitution
ThrowCompletionOr<String> get_substitution(GlobalObject& global_object, Utf16View const& matched, Utf16View const& str, size_t position, Span<Value> captures, Value named_captures, Value replacement)
{
auto replace_string = TRY(replacement.to_utf16_string(global_object));
auto replace_view = replace_string.view();
StringBuilder result;
for (size_t i = 0; i < replace_view.length_in_code_units(); ++i) {
u16 curr = replace_view.code_unit_at(i);
if ((curr != '$') || (i + 1 >= replace_view.length_in_code_units())) {
result.append(curr);
continue;
}
u16 next = replace_view.code_unit_at(i + 1);
if (next == '$') {
result.append('$');
++i;
} else if (next == '&') {
result.append(matched);
++i;
} else if (next == '`') {
auto substring = str.substring_view(0, position);
result.append(substring);
++i;
} else if (next == '\'') {
auto tail_pos = position + matched.length_in_code_units();
if (tail_pos < str.length_in_code_units()) {
auto substring = str.substring_view(tail_pos);
result.append(substring);
}
++i;
} else if (is_ascii_digit(next)) {
bool is_two_digits = (i + 2 < replace_view.length_in_code_units()) && is_ascii_digit(replace_view.code_unit_at(i + 2));
auto capture_postition_string = replace_view.substring_view(i + 1, is_two_digits ? 2 : 1).to_utf8();
auto capture_position = capture_postition_string.to_uint();
if (capture_position.has_value() && (*capture_position > 0) && (*capture_position <= captures.size())) {
auto& value = captures[*capture_position - 1];
if (!value.is_undefined()) {
auto value_string = TRY(value.to_string(global_object));
result.append(value_string);
}
i += is_two_digits ? 2 : 1;
} else {
result.append(curr);
}
} else if (next == '<') {
auto start_position = i + 2;
Optional<size_t> end_position;
for (size_t j = start_position; j < replace_view.length_in_code_units(); ++j) {
if (replace_view.code_unit_at(j) == '>') {
end_position = j;
break;
}
}
if (named_captures.is_undefined() || !end_position.has_value()) {
result.append(curr);
} else {
auto group_name_view = replace_view.substring_view(start_position, *end_position - start_position);
auto group_name = group_name_view.to_utf8(Utf16View::AllowInvalidCodeUnits::Yes);
auto capture = TRY(named_captures.as_object().get(group_name));
if (!capture.is_undefined()) {
auto capture_string = TRY(capture.to_string(global_object));
result.append(capture_string);
}
i = *end_position;
}
} else {
result.append(curr);
}
}
return result.build();
}
}