[m-rev.] for review: fix half of Mantis bug #548

[Julien Fischer]

Hi Zoltan,

On Thu, 10 Feb 2022, Zoltan Somogyi wrote:

> Actually, unless someone objects, I will do a variation of approach 5
> I posted about 20 minutes after the email you replied to. This would
> fall back to implementing the switch using an if-then-else chain if the code
> of an arm that would otherwise have to be shared between two different
> ptags contains both (a) code that will cause the construction of
> an environment, and (b) some block-local variables that would
> then become fields in that environment. Absent either of those factors,
> you can't get a duplicated field name (from this source, that is).

No objections from me to using approach 5 for now.  It's probably the
only reasonable approach for fixing this on the release branch now
anyway.

Julien.