Added undefined property to global scope

[croraf]

Please merge #488 first. Only the last commit is actually for this feature.

Should close #481 when completed

[codecov]

Codecov Report

Merging #501 into master will increase coverage by 0.14%.
The diff coverage is 80.00%.

@@            Coverage Diff             @@
##           master     #501      +/-   ##
==========================================
+ Coverage   67.48%   67.62%   +0.14%     
==========================================
  Files         162      165       +3     
  Lines        9869     9887      +18     
==========================================
+ Hits         6660     6686      +26     
+ Misses       3209     3201       -8     

Impacted Files	Coverage Δ
boa/src/builtins/array/mod.rs	80.09% <0.00%> (-0.20%)	⬇️
boa/src/builtins/mod.rs	25.92% <0.00%> (-1.00%)	⬇️
boa/src/exec/mod.rs	64.70% <ø> (-0.13%)	⬇️
boa/src/syntax/parser/expression/primary/mod.rs	73.68% <ø> (-1.32%)	⬇️
boa/src/exec/identifier/mod.rs	66.66% <60.00%> (-22.23%)	⬇️
boa/src/builtins/undefined/mod.rs	66.66% <66.66%> (ø)
boa/src/builtins/undefined/tests.rs	100.00% <100.00%> (ø)
boa/src/environment/lexical_environment.rs	74.52% <100.00%> (+0.69%)	⬆️
boa/src/exec/operator/mod.rs	66.88% <100.00%> (+0.22%)	⬆️
boa/src/exec/operator/tests.rs	100.00% <100.00%> (ø)
... and 10 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e674da4...50f1381. Read the comment docs.

[croraf]

I've made some changes on how get_binding_value works. I made it to return an Option instead of Value. This is important because we should not return Value::Undefined when the variable (that is node Identifier) is not found. Instead error should be explicitly signaled (for which I'm using Option(None)).

The actual JS error (ReferenceError) is thrown in the execution (run method) of the Identifier node.

There are some other places where get_binding_value is called for some reason. There I just added an unwrap as needed.

[Razican]

It's looking pretty good :) check the comments, it's almost ready!

[croraf]

I've done the replacement. So you can merge if everything else is OK.

Not really sure why there exists the throw_reference_error" function at all? Should it be deleted?

[HalidOdat]

Not really sure why there exists the throw_reference_error" function at all? Should it be deleted?

For a log time we only had the .throw_some_error which return a ResultValue which made throwing errors very easy in builtin functions, but a bit awkward for functions that return Result<NonValue, Value>, like to_number() which returns Result<f64, Value> we had to call unrechable! or use .expect("throw_type_error should always throw") which made this a bit convoluted.

So we added them in #419 to make it easier, we can use the .construct_some_error() in builtin functions, but we have to specify wether it we are throwing with Err() or returning an a Value with Ok(), so the .throw_some_error() are just a convenience functions.

Here we decided to add them: #419 (comment)

[HalidOdat]

Looks perfect to me :)