

FORTRAN-lint QUESTIONNAIRE

This questionnaire suggests specific items to look for when evaluating
FORTRAN-lint.  It doesn't cover all possible issues;  please feel free
to add comments on any areas of concern.

When completed, please email, fax or snail mail to:

    Monty Swaiss

    Cleanscape Software International
    2231 Mora Drive, Suite E
    Mountain View, CA 94040-1548
    (650) 864-9600 (voice)
    (650) 864-9500 (fax)
    E-mail: lint@cleanscape.net

I. Error-Handling

    A. Parser re-synchronization after a syntax-error.

        1. After certain kinds of errors, FORTRAN-lint skips to the end
           of the current statement, then resumes parsing. Does FORTRAN-
           lint seem to skip ahead too far after minor errors?







        2. Does the program print out "false alarm" error-messages caused
           by a small number of actual errors? To the extent possible,
           FORTRAN-lint attempts to print only one message per error in
           the source program. The classic example is the "invalid cont-
           inuation line", which causes some compilers to emit dozens of
           error messages.







    B.  Language Recognition.

        1. Does FORTRAN-lint accept syntax your compiler rejects?







        2. Does FORTRAN-lint flag as "syntax errors" code that your
           compiler accepts?










 C.     Clarity of Errors and Warnings

        1. Are the messages clear and helpful?








II) ANSI Compatibility and Local Compiler Extensions:

        A. Is FORTRAN-lint's support of the Fortran 90 standard adequate?
           If not, what features should be added?










        B. Does FORTRAN-lint adequately cover your FORTRAN compiler's
           language extensions?  If not, please list any unsupported
           extensions.












III)  General Observations:

        A. Does varying the use of switches on the command line create
           any inconsistencies?










        B. Does varying the position of the switches on the command line
           create any inconsistencies?








        C. Are there any inconsistencies when testing FORTRAN-lint in
           global versus local modes (local is the default.)








        D. Are there any discrepancies between the manual and your test
           observations? If so, please document differences.












        E. Examine the output of the Call tree. Are there any missing
           programs or routines from the tree? Is there anything showing
           up in the tree that is not supposed to be there?










        F. Examine the output of the cross reference table. Are there
           any missing symbols that should be displayed? Is there anything
           that should not be in the cross reference table? Are the
           descriptions of all symbols accurate?









IV)  System Information:

        A. What hardware platform and OS level were the tests run under?
           What version of the FORTRAN compiler do you have?



        B. Information about source code:

           1) Number of source files used in testing?

           2) Number of INCLUDE files or libraries included in the above
              code.

           3) Number of source lines?

           4) Is your FORTRAN code compliant with the Fortran 90 stan-
              dard?


V) When providing Cleanscape with source code, please include all
   relevant INCLUDE files or libraries. Source code can be sent via
   E-mail, ftp, 3.5" floppies, 4mm DAT, or 1/4" QIC.

VI) Additional comments:














VII) Any problem areas you would like to see FORTRAN-lint detect?
     Please provide examples.
