Sta update latest 0514

[dsengupta0628]

Latest upstream 5/14
Has support for user defined reporting fields.

Here's an example of how to add a field for an attribute at TCL level.

[get_cell r2] set_attribute orig_name og_r2
sta::make_report_path_attr_field "orig_name" 20
report_checks -fields {orig_name}

From the client code (like mbff) use sta_->makeReportPathField to create the fields

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 2 committers have signed the CLA.

✅ dsengupta0628
❌ jjcherry56
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[gemini-code-assist]

Code Review

This pull request refactors the timing path reporting system to support dynamic and custom fields via a new callback mechanism and map-based field management. It also introduces a LevelizeObserver for cache invalidation during re-levelization and a stringLong utility for robust integer parsing. Feedback identifies a high-severity formatting issue where custom report fields may cause column misalignment, a behavioral discrepancy in whitespace handling between the C++17 and fallback paths of stringLong, and a missing type check in LibertyGroup::findAttrInt that could lead to unsafe type access.

[gemini-code-assist]

The field_case_ and custom fields (via getValue()) are appended to the output line without being passed through reportField. This will cause misalignment in the report columns because these fields won't be padded to their specified width. You should use reportField to ensure consistent formatting and alignment.

      else if (field == field_case_)
        reportField(line_case, field, line);
      else if (field->getValue())
        reportField(field->value(path, this), field, line);

[gemini-code-assist]

There is a behavioral discrepancy between the two implementations of stringLong. std::from_chars (C++17) does not skip leading whitespace and will return an error if the string starts with a space. In contrast, std::strtoll (used in the fallback) skips leading whitespace. This could lead to inconsistent parsing results depending on the compiler and C++ standard library version used. Consider explicitly trimming the string or using a consistent approach for both paths.

[dsengupta0628]

Not changing upstream

[gemini-code-assist]

In LibertyGroup::findAttrInt, the else block assumes that if an attribute value is not a float, it must be a string that can be parsed as a long. However, LibertyAttrValue could potentially hold other types depending on the parser state. Accessing stringValue() without verifying the type (e.g., via isString()) might lead to undefined behavior or assertions if the value is of an unexpected type.

    else if (attr_value.isString()) {
      const std::string &int_str = attr_value.stringValue();
      auto [value1, valid1] = stringLong(int_str);
      value = value1;
      exists = valid1;
      return;
    }

[dsengupta0628]

OpenROAD unit test failure (mbff_orig_name.tcl) is expected as we need OpenROAD side change for the regression to work as well. Check results there- regressions clean there The-OpenROAD-Project/OpenROAD#10437