Changes in column numbers with get-scope

I noticed that in the presence of #includes, the ranges provided by LIGO in get-scope have changed. For instance, suppose I have this:

A.mligo

#include "B.mligo"
#include "C.mligo"
let a = b + c

(note the empty line at the beginning of the file)

B.mligo

let b = 0

C.mligo

let c = 0

Calling get-scope for A.mligo using LIGO 0.52.0 would, for example, say this is the starting point where a is defined:

            "start": {
              "byte": {
                "pos_fname": "test/contracts/find/LIGO-260/A.mligo",
                "pos_lnum": 4,
                "pos_bol": 25,
                "pos_cnum": 29
              },
              "point_num": 29,
              "point_bol": 25
            }

And now by doing pos_cnum - pos_bol + 1, you could retrieve the original column number (5). But with LIGO 0.53.0, this is the output:

            "start": {
              "byte": {
                "pos_fname": "test/contracts/find/LIGO-260/A.mligo",
                "pos_lnum": 4,
                "pos_bol": 1,
                "pos_cnum": 7
              },
              "point_num": 99,
              "point_bol": 92
            }

I don't understand what is the new equation to extract the column number. Initially, it seems like either pos_cnum - pos_bol - 1 or point_num - point_bol - 2, but these equations will fail for other definitions.

Note that even for B.mligo, which has no directives, the output is also different. With 0.52.0:

            "start": {
              "byte": {
                "pos_fname": "test/contracts/find/LIGO-260/B.mligo",
                "pos_lnum": 1,
                "pos_bol": 0,
                "pos_cnum": 4
              },
              "point_num": 4,
              "point_bol": 0
            }

And with 0.53.0:

            "start": {
              "byte": {
                "pos_fname": "test/contracts/find/LIGO-260/B.mligo",
                "pos_lnum": 1,
                "pos_bol": 1,
                "pos_cnum": 4
              },
              "point_num": 90,
              "point_bol": 86
            }

So my question is, is this a regression, or did the way that columns need to be calculated change? If it changed, how should we do it now?