If you have a program like this:

a = 1
degugger
b = 2

When you run it through the interpreter, this would happen before this PR:

From: foo.cr:3:5 <Program>#foo.cr:

    1: a = 1
    2: debugger
 => 3: b = 2

pry> a
1
pry> c = 3
BUG: can't find closured var or local var c (Exception)

That is, you couldn't declare new local variables during a pry session. I think that reduces the usefulness of pry by a lot!

With this PR:

From: foo.cr:3:5 <Program>#foo.cr:

    1: a = 1
    2: debugger
 => 3: b = 2

pry> c = 3
3
pry> c
3
pry> c = "hello!"
"hello!"
pry> c
"hello!"

That is, not only we can declare new local variables, we can also modify their type. This is similar to how it works when you just do crystal i, outside of pry.

Note that you will get an error if you try to change the type of a, because that's already compiled and it would be wrong to change its type:

From: foo.cr:3:5 <Program>#foo.cr:

    1: a = 1
    2: debugger
 => 3: b = 2

pry> a = 'a'
error in line 1
Error: type must be Int32, not (Char | Int32)
pry> a
1
pry> typeof(a)
Int32