[alexanderlinne]

Evaluation results come out in Mono.Cecil loader order, which is deterministic per build but not stable across Cecil or runtime upgrades. Sort the results ordinally when rendering a snapshot so every "Multiple inputs" and "Multiple arguments" block stops depending on that order. The sorted list is also what is handed to ToErrorMessage, so the "Message:" section is stable too; ToErrorMessage itself is production output and stays untouched.

Materialize the evaluation results once in each Assert* method. FormatSnapshot enumerates its argument twice and the All/Any guards add one or two more, so each assertion evaluated its rule up to four times. That is harmless only because every call site currently passes a plain ArchRule, whose Evaluate materializes internally; CombinedArchRule.Evaluate returns a lazy Concat.

AssertOnlyViolations could pass vacuously: ArchRule.Evaluate injects a synthetic failing result when a rule produces no results at all, and the assertion only rejected passing results. A helper lookup that stopped matching would therefore keep the test green. Reject that placeholder unless allowNoResults is set. Only one call site legitimately needs it.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 82.50%. Comparing base (34e4211) to head (8fc416c).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #484   +/-   ##
=======================================
  Coverage   82.50%   82.50%           
=======================================
  Files         259      259           
  Lines       12395    12395           
  Branches     1206     1206           
=======================================
  Hits        10226    10226           
  Misses       1798     1798           
  Partials      371      371           

☔ View full report in Codecov by Harness.
📢 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.