Loaded program test robustness

[bw-solana]

Problem

Some of the current tests in loaded_programs.rs are not robust to different MAX_LOADED_ENTRY_COUNT values. For context, I'm exploring tuning this value down for stress testing as a means of increasing cache churn.

Summary of Changes

1. Make expected entry counts dynamic against max cache entry count.
2. Pull out program_deploy_test_helper helper function to reduce duplicate code. Note that a side effect of this is that we use the same reason for tombstones across all the programs. Previously they were different for each program, but my understanding is stressing this is not the intention of the test (and if it is, it wasn't comprehensive anyhow)

[bw-solana]

Failing due to new security audit. Will wait for #60 to resolve

#60 has been merged and resolved the CI issues

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.8%. Comparing base (3f9a7a5) to head (3fad744).
Report is 3 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##           master      #59     +/-   ##
=========================================
- Coverage    81.8%    81.8%   -0.1%     
=========================================
  Files         837      837             
  Lines      225922   225873     -49     
=========================================
- Hits       184955   184878     -77     
- Misses      40967    40995     +28     

[alessandrod]

Thanks for improving the tests!

Pull out program_deploy_test_helper helper function to reduce duplicate code. Note that a side effect of this is that we use the same reason for tombstones across all the programs. Previously they were different for each program, but my understanding is stressing this is not the intention of the test (and if it is, it wasn't comprehensive anyhow)

Correct, the tombstone type is irrelevant for eviction, so this seems fine.