[m-rev.] for post-commit review: stop putting never-read entries in a map

[Zoltan Somogyi]

For review by anyone. Mostly, I am seeking opinions on whether
we should report any errors in the value of an environment variable
when we discover that this environment variable is intended for Mercury,
or not. The diff adds a big comment outlining the effects of the choices.
Neither choice is obviously superior to the other.

Zoltan.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Log.of2
Type: application/octet-stream
Size: 781 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20251212/35ffdc36/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: DIFF.of2
Type: application/octet-stream
Size: 16615 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20251212/35ffdc36/attachment-0003.obj>