[BUG]: Attempt to resize() an object in a std::optional variable that has not been initialized

[mcordery]

Version

25.02

Which installation method(s) does this occur on?

No response

Describe the bug.

Note this is the same in 25.08.
In the file cpp/tests/sampling/detail/sampling_post_processing_validate.cu, There is a a resize() operation a line 1350:

(*this_type_unique_minor_hops)
            .resize(this_type_unique_minors.size(), handle.get_stream());

The problem here is that when running the test SAMPLING_HETEROGENEOUS_POST_PROCESSING_TEST, the first pass through the loop the std::optional object this_type_unique_minor hops is initialized to std::nullopt (as is this_type_unique_major_hops). When this happens, the attemped resize of the underlying device_uvector seg faults because of the lack of an previously allocated device_uvector. The fix is something like

          if (this_type_unique_minor_hops) {
            .resize(this_type_unique_minors.size(), handle.get_stream());
            (*this_type_unique_minor_hops)
              .resize(this_type_unique_minors.size(), handle.get_stream());
          }

But probably better to be explicit

          if ((*this_type_unique_minor_hops).has_value()) {
            .resize(this_type_unique_minors.size(), handle.get_stream());
            (*this_type_unique_minor_hops)
              .resize(this_type_unique_minors.size(), handle.get_stream());
          }

Minimum reproducible example

Relevant log output

Environment details

Other/Misc.

No response

Code of Conduct

• I agree to follow cuGraph's Code of Conduct
• I have searched the open bugs and have found no duplicates for this bug report

[ChuckHastings]

Thanks for the bug report. Will fix in 25.10.