[m-rev.] for treview: global, local and field vars in the MLDS

[Zoltan Somogyi]

On Sat, 22 Jul 2017 14:49:38 +1000 (AEST), Julien Fischer <jfischer at opturion.com> wrote:
> > The diff passes bootcheck in hlc.gc. For the java and csharp grades, the output
> > of the updated compiler is NOT bit-identical with the output of the previous compiler,
> > as shown by the output of the make_java_csharp_diff tool I committed earlier,
> > which is attached as JAVA_CSHARP_DIFF.
> 
> I'll test the C# and Java backends this evening.

Thank you.

> > There are four kinds of differences.
> >
> > 1. The compiler generates Java/C# code for functions in a different order.
> > To eliminate such differences, I modified both the base and test compilers
> > to sort function definitions before writing them out in mlds_to_{cs,java}.m.
> > Therefore this difference is NOT in JAVA_CSHARP_DIFF. The others are.
> >
> > 2. The new compiler always module-qualifies references to runtime system
> > enum constants that are notionally defined in private_builtin.m.
> 
> As has been noted elsewhere, those enum constants don't really belong
> there.  (I intend to shift them into the runtime at some point.)

The target_prefixes type that this diff added should make that simpler
than it would have been before.
 
> > The old compiler abused the MLDS (treating those constants as if they
> > were global variables) in a way that module qualified them everywhere
> > *except* in private_builtin.java.
> >
> > 3. Almost *all* of the diff part of JAVA_CSHARP_DIFF consists of the
> > following single difference, repeated in lots and lots of times:
> >
> > - /* generic_type */ java.lang.Object closure = null;
> > + /* generic_type */ java.lang.Object closure = closure_arg;
> > cord.Cord_1 conv0_HeadVar__3_3 = null;
> > - closure = closure_arg;
> 
> I suspect that optimization was not occurring in the past precisely
> because of the things that this diff addresses.

So do I, but it may be that the reason is that in this diff, I rewrote part
of the logic of the convert_assignments_into_initializers predicate
in ml_optimize.m, because the old code offended my sensibilities.
It did this by searching for the definition that defines a variable *twice* in a row,
so I deleted the second search after changing the first to remember the location.

Do you have any objection to committing the diff about writing out mmakefile
fragments?

Zoltan.