According to Doug Miller :
In article t, HorneTD
wrote:

That is why the appeals process will presume that the authors of the
code intended it to make sense. It is a basic principle of law that the
writers of the rule in question did not mean to require what cannot be
done nor to forbid what must be done. Since you must apply solder to
the conductors in order to fuse it to the metal the code cannot be
construed to make that impossible. The code language is only intended
to prevent the use of solder as a means of mechanical attachment of a
conductor to a terminal or for the mechanical stability of a splice. A
mechanical splice is "mechanically and electrically secure" without
solder but it will not remain that way if you do not solder it.

That's absurd. If it won't remain that way without solder, it isn't
"mechanically secure without solder".

The CEC has similar wording to the NEC, and as per Knight, soldering _is_
permitted (but he makes quite a point about making sure you do it _right_).

Properly twisted (and undamaged) bare copper inside a box is mechanically
secure and pretty much electrically secure too. For a while at least.
Solder provides permanence.

The whole point behind the NEC and CEC sections is to ensure that you're
not bringing two bare pieces of wire in mid-air into rough proximity,
bridging the gap with a droplet of solder, covering it with scotch tape,
and calling it a day.

That said, soldering is hardly ever done these days simply because having
an iron and taking the time to do it right is generally not worthwhile.

But, there's nothing wrong with it. K&T connections are generally
more reliable than wirenut ones.
--
Chris Lewis, Una confibula non set est
It's not just anyone who gets a Starship Cruiser class named after them.