Regression tester regressed af7473

[tamiko]

[start automated regression testsuite report]

Dear all,

this is the automated regression testsuite reporting a new regression between

• new revision: af74730 regression testsuite results
• old revision: 9bcd4f8 regression testsuite results

Summary:

• 😠 new test failures introduced for configuration GNU-13.1.1-master-autodetection-20230507 on site tester-tos

I have identified the following pull requests as possible candidates:

• Fix a trilinos array type. #16466 (authored-by: @bangerth; merged-by: @kronbichler)
• Simplify Tensor::unroll(). #16467 (authored-by: @bangerth; merged-by: @drwells)

Notes:

• The regression reported above is a subset of the following full set of regressions compared to the baseline:
  • 😞 build warnings for configuration GNU-9.4.0-master-ubuntu-lts-20.04 on site tester-tng
  • 😠 test failures for configuration Clang-16.0.6-master-autodetection-20230507 on site tester-tos
  • 😠 test failures for configuration GNU-13.1.1-master-64bit_indices-20230507 on site tester-tos
  • 😠 test failures for configuration GNU-13.1.1-master-autodetection-20230507 on site tester-tos
  • 😠 test failures for configuration GNU-13.1.1-master-minimal-20230116 on site tester-tos
  • 😠 test failures for configuration GNU-13.1.1-master-standalone_kokkos-20230507 on site tester-tos
  • 😠 test failures for configuration GNU-13.1.1-master-std=c++20-20230626 on site tester-tos
• I will close and unpin this issue automatically once a full run is complete and compares cleanly to the baseline.
• If closed I will reopen the issue if the testsuite run identified an additional regression compared to what I have reported so far.

[end automated regression testsuite report]

[bangerth]

This one confounds me. It's only the "release" mode tests that fail. What might be going wrong here?

[tamiko]

@bangerth I will investigate on the tester directly.

[tamiko]

I am pretty sure this is the same issue as the previous #16463 and related to the DEAL_II_ASSUME macro. We just got a slightly different number of failing tests due to randomness.