[m-rev.] for review: fix and speed up trie string switches

[Peter Wang]

On Sun, 31 Mar 2024 03:30:50 +1100 "Zoltan Somogyi" <zoltan.somogyi at runbox.com> wrote:
> 
> On 2024-03-31 02:56 +11:00 AEDT, "Julien Fischer" <jfischer at opturion.com> wrote:
> > That looks fine otherwise.
> 
> Thanks.
> 
> Do you, or Peter, have an opinion about whether it is ok for deconstruct operations in
> non-canonical mode, when given a string containing ill-formed code unit
> sequences, to return those sequences using octal escapes? The existing
> code throws an exception.

As long as the behaviour is documented, I don't see why it should be
limited to non-canonical mode.

Peter