array_to_subslice3 was not restored following the DST PR

[protz]

7d4d0fd removed the definition of array_to_subslice3 and slice_subslice3

The goal of this phase was to replace occurrences of array_to_subslice (and slice_subslice3) where the range start and end are constant values, with a more optimized macro that does not take so many arguments.

Concretely, the removal of array_to_subslice3 generates a diff like this:

-  Eurydice_slice_copy(Eurydice_array_to_subslice3(uu____0,
-      (size_t)0U,
-      Eurydice_slice_len(slice, uint8_t),
-      uint8_t *),
+  Eurydice_slice_copy(array_to_subslice_3621(uu____0,
+      (
+        KRML_CLITERAL(core_ops_range_Range_08){
+          .start = (size_t)0U,
+          .end = Eurydice_slice_len(slice, uint8_t)
+        }
+      )),

The new code is quite a bit more verbose.

Can we restore this code quality optimization? @Lin23299 thanks!

[protz]

Note that I am open to different code quality optimizations; now that Eurydice_slice is a type abbreviation, one could easily replace this with (Eurydice_dst_XX){ .ptr = uu____0 + 0, .data = Eurydice_slice_len(slice, uint8_t) }

Note that we could also many more code quality improvements now that Eurydice_slice has a definition; for instance, Eurydice_slice_len(e) could be replaced with e.data

[Lin23299]

I've tried to restore this optimization in #313, we can see the diff and discuss there :)