Implement ToggleEvent.source and the request-close command

[Gingeh]

See commit descriptions for spec PRs

[trflynn89]

CI is very unhappy.

[Gingeh]

CI is very unhappy.

yeah no clue what's going wrong, the test passes perfectly on my machine

[tcl3]

I get this crash when running headless-browser -f Text/input/wpt-import/html/semantics/popovers/popover-toggle-source.tentative.html with sanitizers enabled:

/home/tim/repos/ladybird/Libraries/LibGC/Ptr.h:176:17: runtime error: reference binding to null pointer of type 'struct Element'
    #0 0x77630b22934d in GC::Ptr<Web::DOM::Element>::operator*() const /home/tim/repos/ladybird/Libraries/LibGC/Ptr.h:176
    #1 0x77630b22934d in Web::Bindings::ToggleEventPrototype::source_getter(JS::VM&) /home/tim/repos/ladybird/Build/sanitizers/Lagom/Libraries/LibWeb/Bindings/ToggleEventPrototype.cpp:192
    #2 0x7762d963e764 in AK::Function<JS::ThrowCompletionOr<JS::Value> (JS::VM&)>::operator()(JS::VM&) const /home/tim/repos/ladybird/AK/Function.h:148
    #3 0x7762d963e764 in JS::NativeFunction::call() /home/tim/repos/ladybird/Libraries/LibJS/Runtime/NativeFunction.cpp:234
    #4 0x7762d96513df in JS::NativeFunction::internal_call(JS::ExecutionContext&, JS::Value) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/NativeFunction.cpp:163
    #5 0x7762d84f9d1b in JS::call_impl(JS::VM&, JS::FunctionObject&, JS::Value, AK::Span<JS::Value const>) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/AbstractOperations.cpp:96
    #6 0x7762d96b93e1 in JS::ThrowCompletionOr<JS::Value> JS::call<>(JS::VM&, JS::FunctionObject&, JS::Value) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/AbstractOperations.h:131
    #7 0x7762d96b93e1 in JS::Object::internal_get(JS::PropertyKey const&, JS::Value, JS::CacheablePropertyMetadata*, JS::Object::PropertyLookupPhase) const /home/tim/repos/ladybird/Libraries/LibJS/Runtime/Object.cpp:954
    #8 0x7762d96b902a in JS::Object::internal_get(JS::PropertyKey const&, JS::Value, JS::CacheablePropertyMetadata*, JS::Object::PropertyLookupPhase) const /home/tim/repos/ladybird/Libraries/LibJS/Runtime/Object.cpp:911
    #9 0x7762d7e8edd4 in JS::ThrowCompletionOr<JS::Value> JS::Bytecode::get_by_id<(JS::Bytecode::GetByIdMode)0>(JS::VM&, AK::Optional<JS::Bytecode::IdentifierTableIndex>, JS::Bytecode::IdentifierTableIndex, JS::Value, JS::Value, JS::Bytecode::PropertyLookupCache&, JS::Bytecode::Executable const&) /home/tim/repos/ladybird/Libraries/LibJS/Bytecode/Interpreter.cpp:1024
    #10 0x7762d7d43f5d in JS::Bytecode::Op::GetById::execute_impl(JS::Bytecode::Interpreter&) const /home/tim/repos/ladybird/Libraries/LibJS/Bytecode/Interpreter.cpp:2632
    #11 0x7762d7d7c76c in JS::Bytecode::Interpreter::run_bytecode(unsigned long) /home/tim/repos/ladybird/Libraries/LibJS/Bytecode/Interpreter.cpp:611
    #12 0x7762d7d9de4b in JS::Bytecode::Interpreter::run_executable(JS::Bytecode::Executable&, AK::Optional<unsigned long>, JS::Value) /home/tim/repos/ladybird/Libraries/LibJS/Bytecode/Interpreter.cpp:753
    #13 0x7762d8d85ea5 in JS::GeneratorObject::execute(JS::VM&, JS::Completion const&) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/GeneratorObject.cpp:110
    #14 0x7762d8d90293 in JS::GeneratorObject::resume(JS::VM&, JS::Value, AK::Optional<AK::StringView> const&) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/GeneratorObject.cpp:158
    #15 0x7762d87b915b in JS::AsyncFunctionDriverWrapper::continue_async_execution(JS::VM&, JS::Value, bool) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/AsyncFunctionDriverWrapper.cpp:127
    #16 0x7762d87c2e70 in operator() /home/tim/repos/ladybird/Libraries/LibJS/Runtime/AsyncFunctionDriverWrapper.cpp:66
    #17 0x7762d87c2e70 in call /home/tim/repos/ladybird/AK/Function.h:225
    #18 0x7762d963e764 in AK::Function<JS::ThrowCompletionOr<JS::Value> (JS::VM&)>::operator()(JS::VM&) const /home/tim/repos/ladybird/AK/Function.h:148
    #19 0x7762d963e764 in JS::NativeFunction::call() /home/tim/repos/ladybird/Libraries/LibJS/Runtime/NativeFunction.cpp:234
    #20 0x7762d96513df in JS::NativeFunction::internal_call(JS::ExecutionContext&, JS::Value) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/NativeFunction.cpp:163
    #21 0x7762d84f9d1b in JS::call_impl(JS::VM&, JS::FunctionObject&, JS::Value, AK::Span<JS::Value const>) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/AbstractOperations.cpp:96
    #22 0x7762fea1ca38 in JS::call(JS::VM&, JS::FunctionObject&, JS::Value, AK::Span<JS::Value const>) /home/tim/repos/ladybird/Libraries/LibJS/Runtime/AbstractOperations.h:114
    #23 0x7762fea1ca38 in operator() /home/tim/repos/ladybird/Libraries/LibWeb/Bindings/MainThreadVM.cpp:215
    #24 0x7762fea1ca38 in call /home/tim/repos/ladybird/AK/Function.h:225
    #25 0x7762d9890f8f in AK::Function<JS::ThrowCompletionOr<JS::Value> (JS::JobCallback&, JS::Value, AK::Span<JS::Value const>)>::operator()(JS::JobCallback&, JS::Value, AK::Span<JS::Value const>) const /home/tim/repos/ladybird/AK/Function.h:148

I was able to "fix" the issue by making ToggleEvent.source nullable in the IDL file.

[Gingeh]

that sounds like a correct fix to me, because the field is nullable (spec bug?)

[tcl3]

that sounds like a correct fix to me, because the field is nullable

Ah, I'd only did the quickest thing I could think of to resolve the problem. My initial quick reading of the spec led me to believe that source shouldn't ever be null. Looking closer I think that it definitely can be, so I agree that changing this field to be nullable would be correct.

This would also match the definition of CommandEvent.source, which is currently nullable.

[Gingeh]

Should I open a spec issue for the ToggleEvent.source idl definition not being nullable?

[tcl3]

Should I open a spec issue for the ToggleEvent.source idl definition not being nullable?

I think that would be a good idea - then you could add a comment that references the issue to make it clear that we are diverging from the current spec.

[Gingeh]

Good news: The popover source test doesn't crash anymore!
Bad news: Crashed: Text/input/navigation/run-script-before-iframe-initial-navigation.html flake?

[tcl3]

I assume it's a flake? I can't reproduce it locally and the crash only occurs on one of the CI jobs. I've restarted the CI job to see if it happens again.

[Gingeh]

Made a spec issue and added a link

[Gingeh]

yay everything passed!

[Gingeh]

whatwg/html#11360 has already been resolved, so I've removed the corresponding AD-HOC comment