Bug fix for nft::insert_levels(). All tests pass.

[koniksedy]

This PR includes:

• A bug fix for the function mata::nft::insert_level().
  • The main issue was that the result transducer was constructed with an empty delta (on line 387). Consequently, it did not know which number to choose for the new state created using the method mata::nfa::add_state(). This led to the creation of inner states (states with new levels) with the same names as existing states, thereby causing issues with delta.
• Repairs to the expected transducer (on line 638). This transducer contained END_MARKER, even though these markers were only present at synchronization levels.
• Uncommenting the remaining tests for mata::nft::replace_reluctant_literal(). All tests pass.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 74.16%. Comparing base (793521e) to head (be63c6b).

Additional details and impacted files

@@                  Coverage Diff                  @@
##           reluctant_replace     #393      +/-   ##
=====================================================
- Coverage              74.28%   74.16%   -0.13%     
=====================================================
  Files                     43       43              
  Lines                   5456     5457       +1     
  Branches                1215     1215              
=====================================================
- Hits                    4053     4047       -6     
- Misses                   955      966      +11     
+ Partials                 448      444       -4     

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

[Adda0]

LGTM. Sorry, I did not expect you to try to fix the remaining uncommented tests except for the one we talked about. Those other tests were just copy-pasted from the previous test case and not modified yet to try to apply the full replace-all procedure.