>>>>> "Wolfgang" == Wolfgang Jenkner writes:
Wolfgang> Raymond Toy writes:
>> I would think all the places where base-char occur would need to be
>> fixed.
Wolfgang> I think the only problem is f2cl's guessing that the
Wolfgang> array-element-type of a literal string would be BASE-CHAR.
Wolfgang> So a minimal test case would be
Wolfgang> CHARACTER*1 BAR
Wolfgang> PARAMETER (BAR = 'a')
Wolfgang> END
Wolfgang> I actually investigated this a bit but since my ahem... exposure to
Wolfgang> FORTRAN programming is minimal and quite remote I didn't venture to
Wolfgang> propose the following patch, which, perhaps, is not completely insane:
Thanks for the patch. I've been a bit lazy with f2cl lately, but I
think your patch is right.
I'll add your tests cases to the f2cl tests too, once I figure out if
the commented out tests are valid or not. Those are rather hairy
declarations that I always have to look up.
However, I'm a bit reluctant to regenerate all of the converted code
because f2cl has changed since the conversion was done and I don't
want to break anything since there's no test suite in either Fortran
or Maxima for these. But if we start adding some more code for other
functions like incomplete gamma, it might be a good time to regenerate
it all.
Ray