[m-dev.] accepting "any" as "ground" for non-solver types

[Zoltan Somogyi]

On 2024-07-02 15:21 +10:00 AEST, "Julien Fischer" <jfischer at opturion.com> wrote:
>> This is quite unintuitive behavior. The log messages says that this behavior
>> is required to support HAL, but of course HAL is long dead. Is there any
>> *current* reason to retain this behavior? If not, I would very much like to
>> get rid of it.
> 
> I am not aware of any reason for continuing to support that behaviour.

Ok, I have changed it. This requires disabling tests/valid/any_matches_bound,
which makes this a breaking change. Is the break significant enough to
our users to require a note in NEWS.md?

Zoltan.