Avoid pointer cast to time_t from incompatible type

[plexoos]

Resolves #426

[genevb]

If the proposed solution works for builds going forward, I have no objection to it.

[veprbl]

Minor issue:

#10 3672.1 .sl79_gcc485/OBJ/StRoot/StarRoot/TUnixTime.cxx: In member function 'TString TUnixTime::GetGString()':
[27021](https://github.com/star-bnl/star-sw/actions/runs/3347375021/jobs/5545286240#step:3:27024)
#10 3672.1 .sl79_gcc485/OBJ/StRoot/StarRoot/TUnixTime.cxx:135:3: error: 'time_t' is not a member of 'std'
[27022](https://github.com/star-bnl/star-sw/actions/runs/3347375021/jobs/5545286240#step:3:27025)
#10 3672.1    std::time_t utime = fUTime;
[27023](https://github.com/star-bnl/star-sw/actions/runs/3347375021/jobs/5545286240#step:3:27026)
#10 3672.1    ^

[starsdong]

Looks good to me.

[plexoos]

If the proposed solution works for builds going forward, I have no objection to it.

Gene, do you know why we don't see this kind of problems with the 64-bit tests on the farm?

[plexoos]

If the proposed solution works for builds going forward, I have no objection to it.

Gene, do you know why we don't see this kind of problems with the 64-bit tests on the farm?

The answer to my own question is that only a handful of the nightly tests on the farm use 64-bit libraries. By looking carefully at this summary page https://www.star.bnl.gov/devcgi/weekDEVjobStatus.pl one can notice a "_64bit" suffix added to the values in the Path column. Assuming these are the only tests for 64-bit builds, the failed test in CI due to the error reported in #426 is never executed on the farm.

[genevb]

the failed test in CI due to the error reported in #426 is never executed on the farm.

Thanks for doing the investigative work, @plexoos