A new release for GNU Octave

The example given is not representative of the usual way to solve linear systems in Octave.
It is not idiomatic as the usual way would be to use the left division operator, also following the usual notation the matrix is represented with a capital letter:

>> A = [1 2 3; 3 2 1; 2 2 2];
>> b = [2 2 2]';
>> A \ b
warning: matrix singular to machine precision
ans =

0.3333
0.3333
0.3333

As you see Octave is saying that within the used precision the matrix is non-invertible and so the solution is not unique.

Regarding linsolve this is what the first lines have to say about it. (TL;DR linsolve is an advance frontend when used with other options)

>> help linsolve
'linsolve' is a function from the file /home/jamatos/devel/octave/scripts/linear-algebra/linsolve.m

-- X = linsolve (A, B)
-- X = linsolve (A, B, OPTS)
-- [X, R] = linsolve (...)
Solve the linear system 'A*x = b'.

With no options, this function is equivalent to the left division
operator ('x = A \ b') or the matrix-left-divide function
('x = mldivide (A, b)').
....

One can certainly check for the uniqueness of solutions in Octave, but that was not important in this brief example. The returned solution is correct, if not unique.