[DO NOT MERGE] Remove redundant STA calls in RepairDesign

[dsengupta0628]

Summary

Cleans up redundant STA calls in the resizer's RepairDesign, continuing the work started in #10480

Changes

• RepairDesign::performEarlySizingRound / repairDesign: drop redundant searchPreamble() + findAllArrivals(), the explicit findRequireds(drvr->level()+1), the delaysInvalid() call, and the per-net incremental-timing block (ensureLevelized/findDelays/findArrivals). These are consolidated into a single findRequired(drvr) prerequisite in performGainBuffering.
• performGainBuffering: take the driver Vertex* instead of const Pin* (avoids redundant pin→vertex lookups), and add the explicit findRequired(drvr) needed before vertexWorstSlackPath.
• Replace manual Vertex::setSlewAnnotated(false, ...) loops with the Sta::unsetAnnotatedSlew(...) API; use RiseFall::range() instead of explicit rise/fall pairs.
• RepairTargetCollector: route makeExceptionTo and findPathEnds through the Sta API instead of sdc_/search_ directly.

Note: this is the first of two steps. A follow-up will add findRequired inside Sta::vertexWorstSlackPath (the proper place for it), after which the explicit findRequireds/findRequired calls that remain in the resizer can be removed.

Type of Change

• Refactoring

Impact

None

Verification

• I have verified that the local build succeeds (./etc/Build.sh).
• I have run the relevant tests and they pass.
• My code follows the repository's formatting guidelines.
• I have signed my commits (DCO).

[gemini-code-assist]

Code Review

This pull request refactors the timing analysis and repair flow in RepairDesign and RepairTargetCollector. Key changes include updating performGainBuffering to accept a sta::Vertex* instead of a sta::Pin*, adding a prerequisite call to sta_->findRequired(drvr), and removing redundant incremental timing updates and arrival/required search calls. Additionally, manual loops for un-annotating slews are replaced with sta_->unsetAnnotatedSlew, and sdc_->makeExceptionTo and search_->findPathEnds are updated to use sta_ equivalents. There are no review comments, so I have no feedback to provide.

[dsengupta0628]

secure-ci: http://secure-ci:8080/job/SB/job/secure-test-resizer6-apply/

[maliberty]

@codex review

[maliberty]

If the CI is ok I don't have anything to add.

[dsengupta0628]

If the CI is ok I don't have anything to add.

secure CI was fine. Public CI had a small variation in GRT TNS though CF TNS was fine. I created a metric PR The-OpenROAD-Project/OpenROAD-flow-scripts#4308

[dsengupta0628]

The dashboard shows each and every ORFS design is slowed down on all steps compared to the baseline commit in ORFS https://github.com/The-OpenROAD-Project-private/OpenROAD-flow-scripts/commit/c9c22caf9bf9cfe46c5a4236c6ec7e7ae9863cc3 :|
Would like to investigate a bit more, why

[precisionmoon]

Overall code changes look OK. Were there any regression test changes?

[dsengupta0628]

Overall code changes look OK. Were there any regression test changes?

This is not right change as is as runtime degrades. Not closing this just yet in case we decide something else. Please see the attached pdf:
STA_resizing_PR10744.pdf