Re: [xml] HP-UX compilation warnings for 2.6.23

[Rick Jones <rick jones2 hp com>]

> > So, it would appear that a bit of configure manipulation might 
> > suffice, and then on those HP-UX systems with dlopen it will take the 
> > dlopen path rather than the shl_load path.
>
>
> But won't the problem persist on HP-UX versions prior to 11.11?

It might, but I'm not even sure that the HP compilers supported on 11.0 flag the 
warning or not.  11.0 is the only presently "supported by HP" revision of HP-UX 
prior to 11.11 and it probably is not all that long for the HP support world; 
10.20 and earlier are definitely out of support, so the number of people who may 
see it should be dwindling.

While I have no knowledge of plans along those lines, at some point, HP may 
cease supporting shl_mumble and friends, which means that having applications 
use dlopen on HP-UX where it is available would be a good thing, and would be 
the 80% solution I suspect.  It would probably make life easier for those 
applications anyway.

> A possible workaround could be to do something like this (caveat: I
> have not tested it; in fact, I have not looked at the code in question,
> only the above snippet)
>
>   shl_t pointer;
>   rc = shl_findsym(&pointer, name, TYPE_UNDEFINED, symbol);
>   handle = pointer;

would that be handle = pointer or *handle = pointer?  anyway, from the "ironic 
file" I tried the former and got this:

# make xmlmodule.o
        source='xmlmodule.c' object='xmlmodule.o' libtool=no \
        DEPDIR=.deps depmode=none /bin/sh ./depcomp \
        cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include  -D_REENTRANT 
    -g -c xmlmodule.c
"xmlmodule.c", line 279: warning #2550-D: parameter "handle" was set but never
          used
  xmlModulePlatformSymbol(void *handle, const char *name, void **symbol)
                                ^
:)

rick jones