[comp.databases] Problem with Clipper 5.0

crisp@uncecs.edu (Russ Crisp) (12/13/90)

I have a problem with my new clipper 5.0. I have a large
application, consisting of about 65 modules.  When I compile
the main module, clipper 'forgets' to include about half of
them, and I get unresolved symbols at link time.  They are all
accessed by the main routine by DO's, and all are in the same
directory with the main.  Summer 87 has no problem, but 5.0
can't seem to get them all together.  Anyone got a clue here?
Russ Crisp
Systems Analyst
Western Carolina University
internet: crisp@wcuvax3.wcu.edu
bitnet: crisp@wcuvax1.bitnet

tleylan@pegasus.com (Tom Leylan) (12/14/90)

In article <1990Dec13.030106.22861@uncecs.edu> crisp@uncecs.edu (Russ Crisp) writes:
>I have a problem with my new clipper 5.0. I have a large
>application, consisting of about 65 modules.  When I compile
>the main module, clipper 'forgets' to include about half of
>them, and I get unresolved symbols at link time.  They are all
>accessed by the main routine by DO's, and all are in the same
>directory with the main.  Summer 87 has no problem, but 5.0
>can't seem to get them all together.  Anyone got a clue here?
>Russ Crisp
>Systems Analyst

Russ,  there is no reason that 5.0 shouldn't act the same in this regard.
You aren't referencing the DO <filename> with a macro are you ?  Clipper
wouldn't be able to find the file in that case.  You do have them named
.PRG right ?  Again Clipper assumes the extension if haven't told it
anything else.  The files are in the same directory that you are compiling
in ?  If you use a script file the automatic compiling of files is over-
ridden as it would be if you use the /M compiler option.

All that said... may I now try an convince you to avoid that "all" method
of compiling and suggest you use a script file or the make utility.  Besides
creating one giant .OBJ file you circumvent the ability to just compile the
few .PRG files that have changed.

Be happy to discuss your eliminating PROCEDURES altogether and sticking to
the FUNCTION method but that deserves a conversation of its own.

Tom Leylan
Ex-Senior Systems Analyst - Nantucket Corporation