Enable drd and helgrind in the valgrind jobs

[ldorau]

Enable drd and helgrind in the valgrind jobs.

Fixes: #156
Fixes: #157

[igchor]

Do we actually want to run drd and helgrind tests? We already have tsan which cover most (I think) of the things checked by drd/helgrind and tsan has better support for the standard library. Also, it seems that it does not report any false positives for jemalloc/tbb.

[bratpiorka]

@igchor The question is how many of these problems found are real. We have already run TSAN, which reports no problems, so if it turns out that ALL drd/helgrind problems are false positives, then I agree that running TSAN is sufficient - there is no point in maintaining a false-positive-only suppression list. I think the next step should be to merge this PR, close issues 156 + 157 and add a new issues for drd/helgrind results (with a note that we should possibly remove it if it makes sense).

[lukaszstolarczuk]

pls add a link to your execution of this or possibly run this workflow on push trigger to see the output

nevermind, I didn't see your prev. PR enabling Valgrind on PR.

[ldorau]

pls add a link to your execution of this or possibly run this workflow on push trigger to see the output

nevermind, I didn't see your prev. PR enabling Valgrind on PR.

@lukaszstolarczuk So is it OK now ?

[ldorau]

Rebased

[igchor]

I think there is some problem with the script. When I run test_valgrind.sh everything is reported as "PASSED" but when I run valgrind manually (valgrind --tool=memcheck ./test/umf_test-jemalloc_pool) then it reports errors without the suppresions.

[igchor]

Please verify that the script works correctly. Valgrind still fails on my machine.

[ldorau]

@igchor the script works correctly now

[bratpiorka]

@igchor could you confirm?

[igchor]

@igchor could you confirm?

Yes, it looks good now. The tests fail without the suppressions as expected.