Josh Grams wrote:
> PFE doesn't have the FLOATING wordset loaded by default, and also
> has a module which keeps floating-point numbers on the data stack,
> so I've been using it to verify that my new tester works under
> those conditions. Along the way I turned up a couple of bugs.
> When I found the first one, I sent an e-mail to the maintainer
> listed in the README file (Guido Draheim), but didn't get an
> answer. And the Sourceforge CVS statistics page shows no writes
> to the repository in the last 12 months...
No, PFE isn't dead, and my understanding is that Guido is
looking at the changes you requested. Apparently other things
have absorbed most of his energy for quite a while, but I just
heard from him that now he has a bit more time, and is setting
up a system that should be more responsive to requests for
changes.
That's good news for me. PFE has been my primary Forth
environment for several years. One of the strong points of PFE
is the utter ease with which one can implement new, dynamically
loadable libraries of words in universal assembly language, AKA
C. And the ease with which one can write bindings that access
foreign C or Fortran libraries. I don't want to give that up.
> In case anyone cares:
>
> The fpnostack module doesn't know about the FLOATING-STACK
> environment query (should answer 0). This is easy to fix.
Krisha Myneni may well be looking at that as we speak...
> The line numbers for error messages (ABORT etc.) start at zero,
> not at one as most applications do. This is probably also easy to
> fix, but I haven't bothered.
-- David
|
