Use is_trusted bool in insert_shreds() instead manually adjusting root

[steviez]

Problem

The test_duplicate_with_pruned_ancestor test needs to get around a
limitation where the shreds with a parent older than the latest root are
discarded. The previous approach manually adjusted the root value in the
blockstore; this is not ideal in that it is fiddling with the inner
working of Blockstore.

Summary of Changes

So, use the is_trusted argument in Blockstore::insert_shreds(); setting
is_trusted=true bypasses the sanity checks (including the parent >=
latest root check).

[AshwinSekar]

thanks for this, a much better solution than my hack! just a minor nit on a comment

[codecov]

Codecov Report

Merging #34010 (142e35c) into master (28e08ac) will increase coverage by 0.0%.
Report is 3 commits behind head on master.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master   #34010   +/-   ##
=======================================
  Coverage    81.9%    81.9%           
=======================================
  Files         811      811           
  Lines      219427   219424    -3     
=======================================
+ Hits       179769   179771    +2     
+ Misses      39658    39653    -5     

[steviez]

I was initially a little concerned in that one of the local-cluster CI steps took 3 retries to succeed. However, looking at the logs for those 3 failures, all 3 failed at the same spot on a different test:

thread 'test_optimistic_confirmation_violation_without_tower' panicked at local-cluster/tests/local_cluster.rs:3382:13:

Violation expected because of removed persisted tower!

test test_optimistic_confirmation_violation_without_tower has been running for over 60 seconds

test test_optimistic_confirmation_violation_without_tower ... FAILED

This is not the test whose behavior I altered; test_optimistic_confirmation_violation_without_tower does call copy_block() as well, but it has a value of false set which matches the pre-existing behavior. Namely, this false matches the previously hard-coded false in copy_blocks().

So, it would appear that this test was previously flaky, and my PR does not alter the flaky test so I think I can ok to submit. Do you agree with my conclusion @AshwinSekar; also CC @yihau as I know you've been keeping an eye on flaky tests

[AshwinSekar]

yeah that's a known flaky test. i was unable to reproduce it locally and then it quickly got buried in my todo. I'll take a look again at it next week #29221