How about as a solution that we add a VM flag to enable the backwards
compatibility for MacOS Metal?   Make the default what works quickly
on the current systems but, if someone has a problem it is possible to
return to the old behaviour.

Then we can see how often it seems needed and possibly remove this in
the future.

It is true that causing problems on old images is a problem and we
have had a somewhat recent discussion with Ted? about this where some
of his old images no longer run.

But... since we are talking about MacOS the bigger problem with MacOS
is Apple's decision to disallow running 32 bit code, and, our 32 bit
images requiring a 32 bit VM.    And that was Teds problem with his
old images.

We are a small group and there are limits as to what we can do.  I
think that we should try to work well on current systems even if it
means that older systems stop working.  It is sad, I have some of
those older systems, but that is life..



On 2022-05-06T21:49:52.000+02:00, Eliot Miranda
***@***.***> wrote:

> On Fri, May 6, 2022 at 2:23 AM Marcel Taeumel ***@***.***>
> wrote:

>>  If the VM is hardware, then Squeak is not an application but an
> operating
>>  system. And not all operating systems run on new hardware. They
> might need
>>  patches. And that's okay. They need patches to make the actual
> applications
>>  in them work.
>>  I think that you current perspective on compatibility between VM
> and Image
>>  is unnecessarily restrictive. The image is not an application but
> an
>>  operating system.

> An image may be include operating system (e.g. a trunk dev image),
> but is
> more. An image may be a deployed application (e.g. stripped of the
> IDE, or
> including a restricted IDE, e.g. Scratch). Vuewing an image as "an
> operating system" is vague. Do OS's include all the apps that can
> run on
> them? Not for me. So an image is potentially much more, and
> potentially
> much less, than an OS. In any case, it is something that expects
> certain
> semantics of the VM upon which it runs, and it is unreasonable and
> unacceptabvle for that VM to arbitrarily change its semantics,
> unless it is
> fixing bugs.

> —
>>  Reply to this email directly, view it on GitHub
> <https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/627#issuecomment-1119423119>;,
>>  or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/ADY5VUG3QJHIES2EPP7HX4DVITQKRANCNFSM5URRNFFQ>;
>>  .
>>  You are receiving this because you commented.Message ID:
>>  ***@***.***>

> -- 
> _,,,^..^,,,_
> best, Eliot

> —
> Reply to this email directly, view it on GitHub
> [https://github.com/OpenSmalltalk/opensmalltalk-vm/issues/627#issuecomment-1119961593],
> or unsubscribe
> [https://github.com/notifications/unsubscribe-auth/ACEXJIWH2IR3B6KWXFQ6Y5LVIVZWBANCNFSM5URRNFFQ].
> You are receiving this because you are subscribed to this
> thread.Message ID:

Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.Message ID: <OpenSmalltalk/opensmalltalk-vm/issues/627/1120170170@github.com>