A bug listed here means that f2perl outputs some translated code that runs but, without warning, gives an incorrect result compared to Fortran.
See the README file for a list of other limitations of f2perl which are not considered bugs.
An integer divide which occurs within a statement will probably be
converted to a perl (floating point) divide, which will produce an
incorrect result. For example see the test file int.f. Please check
your results carefully, and add int functions around such divides.
This problem can be fixed with some additional coding.
When an array is passed to a subroutine, as in
dimension a(10) call sub1(a(2))
there is an ambiguity as to whether an offset array should be passed, or just a scalar. It doesn't matter to the calling routine in Fortran, but it does in perl. At present, an offset array is assumed. See file protobug.f for an illustration. A future solution to this problem may be to do two passes through the source, with the first pass determining the appropriate prototypes for call parameters.
If the number of items in a formatted output list exceeds the number of format descriptors, Fortran will still output all items by repeating the last group of descriptors. F2perl currently translates format descriptors to printf-style formats, which stop when the format descriptor list is exhausted. For example, see the file format1.f.