1. 22 Jul, 2019 1 commit
  2. 03 Jul, 2019 1 commit
  3. 28 Jun, 2019 1 commit
  4. 05 Jun, 2019 1 commit
  5. 16 Apr, 2019 1 commit
  6. 12 Apr, 2019 1 commit
  7. 07 Apr, 2019 1 commit
  8. 05 Apr, 2019 1 commit
    • GILLES Sebastien's avatar
      #1398 Change teat utility used from Catch2 to Boost.Test. · 28057a71
      GILLES Sebastien authored
      The reason for the change was partly bug #1363: tests were failing for macOS / shared library / release mode solely due to an internal error of Catch2.
      
      Now we need to have Boost Test installed (as usual ThirdPartyCompilationLibrary facility provides it).
      
      I also used the modifications to refresh the way tests were written:
      
      - Fixture are now handled in a much better way; in peculiar some tests no longer required a fake model and its equally fake Lua file to proceed.
      - The model tests have been refined to provide more information in case of failure (the complete path of the files that might not be equal are now directly given).
      - Mpi fixture is now working completely: you may define several test suites in an executable without having it fail (previously it was the case only for the full-fledged model fixture).
      
      This commit is unusually big for me, but that is due to the fact with macro everywhere in both Boost.Test and Catch2 I had rather not used them at the same time.
      28057a71
  9. 01 Apr, 2019 1 commit
  10. 26 Mar, 2019 1 commit
  11. 19 Mar, 2019 3 commits
  12. 23 Nov, 2018 10 commits
  13. 09 Nov, 2018 1 commit
  14. 08 Nov, 2018 3 commits
  15. 15 Oct, 2018 2 commits
  16. 05 Oct, 2018 4 commits
  17. 07 Sep, 2018 3 commits
  18. 31 Aug, 2018 1 commit
  19. 30 Aug, 2018 1 commit
    • GILLES Sebastien's avatar
      #1335 Remove entirely the macro MOREFEM_CHECK_NAN_AND_INF; the few tests... · 58de6df4
      GILLES Sebastien authored
      #1335 Remove entirely the macro MOREFEM_CHECK_NAN_AND_INF; the few tests checked were actually too constraining for some models (hyperelastic models may lead to infinite in some Newton steps, in which case the Newton diverges ad another attempt might be made).
      
      I put in Utilities/ThirdParty new functions to check whether a floating point, a Seldon vector or a Seldon matrix includes some inf or nan; it is not currently in use but a unit test has been added to ensure their correct behaviour.
      58de6df4
  20. 21 Aug, 2018 1 commit
  21. 24 Jul, 2018 1 commit