to_vrs returns valid results when hgvs position is greater than ref sequence length

I'm not exactly sure what is happening here, but it seems wrong and is easily reproducible. It appears that when the `same as reference` hgvs syntax is used as shown below, that you can put a position in that is greater than the length of the sequence and it will still return a result with a state of `state: { type="LiteralSequenceExprssion",value=""}`. You can test it with the following expression...

`NM_000412.5:c.1930=`

The transcript `NM_000412.5` is `1947` residues long, but its CDS region is`19..1596`, so I'm assuming the `c.1929` is really be reference position `1929+(19-1)` or `1947` thus pointing to the last residue in the sequence. So, I would assume anything greater than c.1929 would fail. But as I write this out I'm thinking that `c.1929` should really fail since the coding sequence is really goes from the ref seq residue `19` through `1596`. This means the c. positions would go from `c.1 = 19` to `c.1577 = (1596 - 19)`. If this tracks then I would assume any c. position greater than `c.1577` should fail. Of course, if someone used the hgvs positional syntax to reference sequences further into the 3\`utr region like `c.*200` which would indicate an additional 200 residues into the 3\`utr region past the stop codon of the coding sequence.

In this transcript the last position that is valid using c. nomenclature would be `1947 base ref seq len - 1596 last cds position = 351` or `NM_000412.5:c.*351=`.