This is pretty heavy and unoptimized, but it will do the trick for now.
Basically, Heap now has a HashTable<HandleImpl*> and you can call
JS::make_handle(T*) to construct a Handle<T> that guarantees that the
pointee will always survive GC until the Handle<T> is destroyed.
Now that Interpreter keeps all arguments in the CallFrame stack, we can
just pass a const-reference to the CallFrame's argument vector to each
function handler (instead of copying it.)
This patch adds a CallFrame stack to Interpreter, which keeps track of
the "this" value and all argument values passed in function calls.
Interpreter::gather_roots() scans the call stack, making sure that all
argument values get marked. :^)
We now scan the stack and CPU registers for potential pointers into the
GC heap, and include any valid Cell pointers in the set of roots.
This works pretty well but we'll also need to solve marking of things
passed to native functions, since those are currently in Vector<Value>
and the Vector storage is on the heap (not scanned.)
This commits makes effort towards tolerating some of javascript's quirks
when it comes to its type system, note that the interpreter's way of
handling type coercion is still not mature at all, for example, we still
have to implement NaN instead of just crashing when trying to parse a
string and failing.
Instead of every NativeFunction callback having to ask the Interpreter
for the current "this" value and then converting it to an Object etc,
just pass "this" as an Object* directly.
This patch adds NativeProperty, which can be used to implement object
properties that have C++ getters and/or setters.
Use this to move String.prototype.length to its correct place. :^)
To make sure that everyone has the same instance of StringPrototype,
hang a global prototype off of the Interpreter that can be fetched
when constructing new StringObjects.
This patch adds String.prototype.charAt() to demonstrate that prototype
property lookup works, and that you can call a prototype function on an
object, and it will do what you expect. :^)
Object will now traverse up the prototype chain when doing a get().
When a function is called on an object, that object will now also be
the "this" value inside the function. This stuff is probably not very
correct, but we will improve things as we go! :^)
We now evaluate for loops in their own scope if their init statement is
a lexical declaration.
Evaluating for loops in their own scope allow us to obtain expected
behaviour, which means for example, that the block-scoped variables
declared in a for statement will be limited to the scope of the for
loop's body and statement and not to that of the current scope (i.e the
one where the for statement was made)
Previously, we were allowing the redeclaration of a variable with `let`
or `const` if it was declared initially using `var`, we should not
tolerate any form of variable redeclaration using let/const.
This patch makes all HeapBlock allocations aligned to their block size,
enabling us to find the HeapBlock* for a given Cell* by simply masking
bits off of the cell address.
Use this to make a simple Heap& getter for Cell, which lets us avoid
plumbing the Heap& everywhere.
Both types of functions are now Function and implement calling via:
virtual Value call(Interpreter&, Vector<Value> arguments);
This removes the need for CallExpression::execute() to care about which
kind of function it's calling. :^)
