Re: [bug-cflow] cflow: static functions in inverted trees

[Sergey Poznyakoff]

Hi Chris,

> I've been using cflow to analyse our source trees (great tool by the
> way, thanks!), and spotted one detail that surprised me. We're most
> interested in inverted trees, where each callee lists its callers. If
> you do that, static functions aren't included in the original list of
> callees so we don't get their details. Is there a reason for that?

I'm afraid I don't quite follow you.  Static functions are included by
default, for both direct and inverted graphs.  A test run on the sample
wc.c source shows: 

$ cflow -r cflow-1.5/doc/wc.c | head
COUNT():
    getword() <int getword (FILE *fp) at ../doc/wc.c:78>:
        counter() <void counter (char *file) at ../doc/wc.c:108>:
            main() <int main (int argc, char **argv) at ../doc/wc.c:127>
counter() <void counter (char *file) at ../doc/wc.c:108>:
    main() <int main (int argc, char **argv) at ../doc/wc.c:127>
exit():
    error_print() <void error_print (int perr, char *fmt, va_list ap) at 
../doc/wc.c:22>:
        errf() <void errf (char *fmt, ...) at ../doc/wc.c:34>:
            main() <int main (int argc, char **argv) at ../doc/wc.c:127>

Notice the errf function in the next to last line.  It is defined in
wc.c as

static void
errf (char *fmt, ...)
{

Am I missing something?

Regards,
Sergey