Implement expression group transcription for conformance DSL.

[nirosys]

Issue #, if available: #935

Description of changes:
This PR adds support for transcribing expression groups in the toplevel clause of the DSL, similar to expression invocations.

Additionally, a small change was also made to the mactab clause to automatically include the current unnamed module when re-defining it in order to keep $ion imported.

Example usage:

(ion_1_1
   (toplevel ('#$:values' ('#$:' 0 1 2)))
   (produces 0 1 2)
)

---

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 78.58%. Comparing base (138ed13) to head (ee4a692).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #953      +/-   ##
==========================================
- Coverage   78.58%   78.58%   -0.01%     
==========================================
  Files         138      138              
  Lines       35210    35210              
  Branches    35210    35210              
==========================================
- Hits        27671    27670       -1     
  Misses       5475     5475              
- Partials     2064     2065       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nirosys]

Coming back to this. After merging with main, to resolve conflicts, some of the conformance dsl tests are failing.

Post-merge the toplevel fragment now handles pre-populating macros that are defined in the toplevel's sub-elements. Unfortunately those elements are then serialized to the input doc which re-defines the macros, resulting in decoding errors.

For example one of the tests uses the following 2 fragments:

   (mactab (macro m (v!) (%v)))
   (:m 1)

The test is to ensure that we don't add new IVMs between the fragments, which was an issue previously.

When these fragments are serialized, it produces (formatting added):

$ion_1_1
$ion::(module _ (symbol_table _ ) (macro_table _ (macro m (v ) ('%' v ) ) ) )
$ion::(module _ (macro_table _ (macro m (v '!' ) ('%' v ) ) ) )
(:m 1)

It appears the change made to the top-level fragment is adding the macros, but they're being written out in addition to the toplevel elements that are defining them as well. This results in an error when deserializing:

Decoding(DecodingError { description: "macro named 'm' already exists", position: None })

It might be better to walk the toplevel elements looking for encoding directives and register them as we find them so that we can skip serializing them with the rest of the elements.

Looking to see how involved that will be.