The log does not show the correct severity level which makes it difficult to identify important messages. For example, the output of a successful run of testConfigOutput contains
[2026-02-09 16:20:55.467942] [0x00007e33aba36b80] [info] Accessing the ModelComponent::getStore() after it was destroyed. A new store will be constructed. In a test with multiple cases this is the desired behaviour.
[2026-02-09 16:20:55.468317] [0x00007e33aba36b80] [info] ConfigOutput: No field with the name "top_melt" was found.
[2026-02-09 16:20:55.468868] [0x00007e33aba36b80] [info] ConfigOutput: Outputting 4 fields to diag0111.nc at 2020-01-11T00:00:00Z
Everything is tagged as [info] even though the first two messages are actually warnings. This is the expected behavior since we use custom severity levels in boost::log. According to the docs one has to define a suitable operator<< for the severity type in order to get customized formatting. However, my initial attempt at this did not work and maybe it would be better to fully customize the outputs. The second argument [0x00007e33aba36b80] of the standard format is also not helpful with our current logging setup.
The log does not show the correct severity level which makes it difficult to identify important messages. For example, the output of a successful run of
testConfigOutputcontainsEverything is tagged as [info] even though the first two messages are actually warnings. This is the expected behavior since we use custom severity levels in
boost::log. According to the docs one has to define a suitableoperator<<for the severity type in order to get customized formatting. However, my initial attempt at this did not work and maybe it would be better to fully customize the outputs. The second argument[0x00007e33aba36b80]of the standard format is also not helpful with our current logging setup.