checklist

  1. Testcases, part 1
    Check your testcases accurately (rules above 3.4.1) and show the results in the FASTEST-Group meeting.
  2. Memory
    1. No local fields with parameters allocated (see 3.4.3).
    2. All fields allocated via #include dim&sstarf#star; are passed through the parameter lists of the subroutines (from fhp, over fmg3d and all other subroutines downto the actual subroutine).
    3. New memory is only allocated for new models, if they are switched of (in *.md memory won't be allocated. Check it with the command size.
  3. Residuals
    If there is nothing changed in the numerical procedures or approximations, the residuals don't have to change. Check it with the Tkanal for the laminar cases. (Additionaly a check for your used models could be reasonable in regard to your changes.)
  4. Compiler-Check/Run time Check
    There exists compilers with a hard FORTRAN check like portland compilers or IBM-compilers. The gnu-compiler is very usefull to find segmentation faults. Compiler check on IBM is obligatory, run time check on gnu-compiler too. The compiler check includes the execution of the program!
  5. CVS-Check
  6. Testcases, part 2
    Check your testcases accurately again (using the new checked out version). Insert it in to the database [*].

  7. Documentation
    Extend the documentation:
    1. comments in the code
    2. developers guide
    3. users guide.
    The rules for comments in the code is described in the chapter Code Design3.3. Describe new variables and variables you ascertained in the developers guide, same to subroutines. Describe your improvement and the new input parameters in the user guide. (CVS-informations about the user guide in 3.1.1).