Re: [open-cobol-list] Proposed addition to OpenCOBOL: an object module generator

[David Essex]

Jeff Chimene wrote:

> ...
> That's certainly the approach I'd take if I
> was starting from scratch. 

Well, actually NOT from scratch.
You have a scanner/parser, run-time, syntax tree.
I would say about 40-50% complete. Over optimistic ? Maybe.

In any case I think it can be done, with the proper sponsor, financing.


> The biggest hurdle I see is official incorporation into GCC.
> Perhaps that's not as big a hurdle as it seems.

Not likely to happen any time soon.
But then again, other compilers use the GCC back-end.
Some, such as GCC-Pascal, eventually became part of GCC (FSF supported
projects).


> On the debugging side, here is another possible approach. It's
> basically the one that's appeared in various guises in this thread:
> http://www.mono-project.com/Debugging
>
> Ignoring the JIT stuff, it would seem that cobc is well on the way to
> supporting the necesary call-backs as demonstrated by those Python
> data dump routines.
>
> Given the wealth of scripting languages supported by OC, it's really a
> matter of designing and implementing useful COBRTL routines for them
> to call.
>
> COBOL_OC_DUMP is certainly useful.

Which ever approach chosen, you still have to find a way to pass
debugging information (COBOL->C->binary-object).
No small task, I don't think.

Perhaps the simplest possible approach is to add 'READY/RESET TRACE'
COBOL extension. In combination with 'DISPLAY', it is a good tool for
debugging classic type COBOL programs.


Cheers.