Re: Strange Porting Problem CVF to IVF, requesting help and ideas.

On 2006-09-27 15:55:53 -0300, glen herrmannsfeldt <gah@xxxxxxxxxxxxxxxxxxxxxxx> said:

Steve Lionel <steve.lionel@xxxxxxxxx> wrote:

Array bounds checking is good but doesn't help if you use (*)
dimensioned arrays. Argument type mismatches can also get you into
trouble - in Intel Fortran, add the options /gen_interface and
/warn:interface to help catch these.

The suggestion was that arrays in other COMMON blocks
were overwriting the desired variable. I don't believe (*)
is allowed in COMMON arrays.
But yes, if a COMMON array were passed as a subprogram
argument it could be referenced through a (*) dimension and
not be bounds checked.

-- glen

Salford (Siverfrost) FTN95 uses a bunch of hidden arguements to
address this problem. They go to some trouble to make thier
checked and unchecked code interoperable. The level of complication
required leads to stability issues. It is great when it works
and tends to work better on the F77 subset. There is an older
FTN77 that is very solid. I am not clear on the state of
Lahey or NAG as I try to not tempt fate more than neccessary.




.

Relevant Pages

  • Re: DIM vs. Dynamic
    ... not matter if the equates are simply an include of a standard BP schema item ... dynamic arrays that got me thinking. ... 100,000 assignments were finished in the dynamic array case in only ... The difference between dynamic and dimensioned arrays ...
    (comp.databases.pick)
  • Re: Total order on floating point numbers
    ... glen herrmannsfeldt writes: ... |>> In C89, what you did was about the best that could be done; ... |> I used it with arrays of pointers to arrays. ...
    (comp.arch.arithmetic)
  • Re: Percentage of the occurance of a word in a date range
    ... I'm still having trouble (probably because I don't ... have a good grasp of arrays). ... I tweaked it a little because I am referencing ...
    (microsoft.public.excel.worksheet.functions)
  • DIM vs. Dynamic
    ... The previous post on EQU had some discussion of dimensioned arrays vs. ... 100,000 assignments were finished in the dynamic array case in only ... process that runs *way** too long, I'm going with dynamic arrays only ...
    (comp.databases.pick)
  • Re: Intel announces 64-bit extensions
    ... > I see compilers as a mean to save me trouble, ... >> necessarily a lot slower. ... > remotely elegant or just leaning toward the KISS side. ... But, as I said, arrays seem to be problematic. ...
    (borland.public.delphi.non-technical)