The debugger returns the wrong `coor` for atoms

[PEZ]

The debug response coordinates for the dereferenced atom contains an extra 1, throwing the client out of whack. For the repro code below the coor is returned as [3 1], where it probably should just be [3]. At least that would make Calva happy.

A minimal repro project: https://github.com/PEZ/cider-nrepl-debugger-issues, thanks @javierrweiss!

Issue reported on Calva:

• https://github.com/BetterThanTomorrow/calva/issues/2501

Steps to reproduce the problem

In CIDER or Calva:

1. Evaluate the top level forms here:

   (def !a (atom 0))
   
   #dbg
   (defn boom
     []
     @!a)
   
   (comment
     (boom)
     )

2. Evaluate (boom)

Expected behavior

The debugger should stop showing the value 0 from the @!a expression.

Actual behavior

Things go BOOM. A bit different symptoms in Calva and CIDER.

Environment & Version information

cider-nrepl version

0.47.1

Java version

20

Operating system

MacOS

[PEZ]

It's been brought to my attention that [3 1] is indeed the right coor for the dereffed atom. This needs to be fixed in Calva instead.

[bbatsov]

A problem that solves itself! My favorite kind! 😄

[vemv]

Let us know if the documentation could be improved somewhere

Cheers - V