Fix depth-preserving rewrite

[MyskYko]

regarding the issue #687
there was a wrong inequality.
also thinking to clean up code a bit: depth computation is duplicated (depth view already does it) and seemingly incomplete (it only updates two TFO levels for each replacement, which is probably okay for most cases, followed by one TFI level check per iteration, but it might not work when a target node has more than 2 levels of fanouts that are already visited or new, so the follow-up one TFI level check fetches old level).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.79%. Comparing base (fa8513a) to head (35a3fe6).
⚠️ Report is 5 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #690      +/-   ##
==========================================
- Coverage   83.83%   83.79%   -0.04%     
==========================================
  Files         189      189              
  Lines       29888    29034     -854     
==========================================
- Hits        25056    24329     -727     
+ Misses       4832     4705     -127     

☔ 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.

[MyskYko]

fixed another bug in level evaluation, where visited nodes now return correct depth rather than 0.

[MyskYko]

Will create another PR for cleanup...