In continuation of #537 (solved by #721), we now have the following structure:
elastic-tube-1d/
...
reference-results/
fluid-cpp_solid-cpp.tar.gz -> in LFS
fluid-cpp_solid-python.tar.gz -> in LFS
fluid-python_solid-python.tar.gz -> in LFS
reference_results.metadata -> text, directly in the repository
Since we now can generate reference results for individual test suites (added by #728), the reference_results.metadata is updated by the latest run, without necessarily corresponding to all of the archives in this directory.
At the same time, the text is a bit useless without the archives (in LFS), adds noise to the Git history, and often pops up in code searches.
We should break this file into one per archive, and put it into the respective archive, ending up with:
elastic-tube-1d/
...
reference-results/
fluid-cpp_solid-cpp.tar.gz -> in LFS, including a reference_results.metadata
fluid-cpp_solid-python.tar.gz -> in LFS, including a reference_results.metadata
fluid-python_solid-python.tar.gz -> in LFS, including a reference_results.metadata
In continuation of #537 (solved by #721), we now have the following structure:
Since we now can generate reference results for individual test suites (added by #728), the
reference_results.metadatais updated by the latest run, without necessarily corresponding to all of the archives in this directory.At the same time, the text is a bit useless without the archives (in LFS), adds noise to the Git history, and often pops up in code searches.
We should break this file into one per archive, and put it into the respective archive, ending up with: