[m-rev.] for review: string trie switches for the LLDS, non-lookup edition

[Zoltan Somogyi]

For review by anyone. I would appreciate a review
by tomorrow evening, because I want to commit this before
starting work on the lookup version.

Besides a review of the meat of the diff, I am looking for feedback
on two peripheral issues.

One is that this diff exports a test for the use of utf8 by
the platform the compiler is running on. This works for
what this diff needs, but I wonder whether people would
prefer that I replace that with a predicate in (say) globals.m
that is implemented by three foreign procs, each of which
returns the compilation_target of the language of the
foreign_proc.

The other is: does anyone know why --no-inlining does not
actually inhibit inlining? Was this a deliberate choice, or just
an accident?

Zoltan.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Log.strsw1
Type: application/octet-stream
Size: 4276 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20240325/2ab6a9d1/attachment-0002.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: DIFF.strsw1
Type: application/octet-stream
Size: 68070 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20240325/2ab6a9d1/attachment-0003.obj>