PS: Add field flow

[MathiasVP]

This PR implements field-based dataflow (i.e., "field flow") for Powershell.

Field flow is what allows us to get flow in examples such as:

$x.f = Source
Sink $x.f

While it may look innocent, field flow is extremely powerful, and extremely hard to get to perform well. Luckily, all the complications are hidden away inside the shared dataflow library, so all we need to do is:

• Make sure there exists DataFlow::Nodes to represent the notion of "a dataflow node where we have just stored data into". These nodes are known as "post-update nodes".
• Tell the dataflow library what a write into a field looks like. This is the job of storeStep.
• Tell the dataflow library what a read of a field looks like. This is the job of readStep.

Consider an example such as:

$x.f = Source # line 1
$y = $x # line 2
Sink $y.f # line 3

operationally, what happens is that:

• There is a storeStep from Source to the post-update node for $x (which is printed as [post] $x). Internally, the data flow "remembers" that there has been a write to f when it generates PathNodes (that's the part that's extremely hard to get to perform well). When viewing paths, this will be printed as something like [post] $x [f]. The [f] part is known as the "access path" and can contain up to 5 entries (which represents tracking a value that's been stored into 5 nested structs/classes).
• There's a flow step from [post] $x on line 1 to $x on line 2 (we get this from SSA).
• There's a flow step from $x to $y
• There's a readStep from $y to $y.f that reads from y. Because the dataflow library "remembers" that there was a previous write to f it will "pop off" f from the access path.