[m-rev.] for review: centralize recognition of sharing for tag switches

[Zoltan Somogyi]

For review by anyone. Peter reviewed my last diff to
ml_tag_switch.m, so he may also want to review the diff
that replaces that one :-(

This diff adds rather a lot of ZZZs, which I intend to leave
in the diff to be committed. I intend to address and then remove
those ZZZs over about the next week. However, I intend to
first break up switch_util.m, because the changes I made to it
were made more complicated by the presence in it
of code that is relevant only to *other* kinds of switches.
Are there any objections?

Many of the ZZZs are about possible speed improvements.
I expect that my laptop does way too much thermal throttling
for it to be able to reliably measure any resulting speed changes,
especially in this heat :-( Does anyone have any machines
suitable for benchmarking?

Zoltan.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Log.tsw
Type: application/octet-stream
Size: 2152 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20240311/3e7f4931/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: DIFF.tsw
Type: application/octet-stream
Size: 123586 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20240311/3e7f4931/attachment-0003.obj>