[petsc-dev] PETSC_USE_SOCKET_VIEWER macro - controlled by configure or not

[Matthew Knepley]

On Tue, Jan 12, 2010 at 9:05 PM, <Kevin.Buckley at ecs.vuw.ac.nz> wrote:

>
> >> I am tying to help a researcher here build his application against
> >> a PETSc (3.0.0 p9) installtion at another site that seems to have
> >> undefined references to:
> >>
> >> PETSC_VIEWER_SOCKET_
> >
> > Is this unresolved symbol from user code?
>
> No.
>
> They are all from within PETSc itself.
>
> Specifically, the various PETSc solvers that have the
>
> #ifdef(PETSC_USE_SOCKET_VIEWER )
>
> block I also forgot to mention in the body but alluded to in the
> subject. (You see, I did go digging in the code and libs!)
>
> To be a bit more verbose, on the machine I have no control over,
> other than to feed info to the installer there:
>
> for f in *.a ;do echo $f ; nm $f | grep PETSC_VIEWER_SOCKET_ ; done
> ...
> libpetsc.a
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
> ...
>
> and so on for other PETSC libs.
>
> On my machine there's obvioulsy a fix up for that reference here:
>
> ...
> libpetsc.a
> 00001a4a T PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
>         U PETSC_VIEWER_SOCKET_
> ...
>
> What I could not discern from my inspection of the code so far
> was how the actual instatiation of the fix up for those references
> might have been binned on the other box, whilst the (U)ndefined
> references themselves still appear to have been compiled into the
> PETSc installation.
>
> I was expecting, once I had found it that is, PETSC_USE_SOCKET_VIEWER
> and a corresponding configure ability to turn it on/off, might control
> everything.
>

This was a bug in the prior release. The fix is to remove the
PETSC_HAVE_SOCKET
lines from $PETSC_ARCH/include/petscconf.h and rebuild. Should not happen in
the
new release of configure. Someone needs to do this on that machine.

  Matt


> Then I found the reference to tweaking it by hand on, admittedly
> an not quite run of the mill, platform.
>
> Hence my original question.
>
> --
> Kevin M. Buckley                                  Room:  CO327
> School of Engineering and                         Phone: +64 4 463 5971
>  Computer Science
> Victoria University of Wellington
> New Zealand
>
>


-- 
What most experimenters take for granted before they begin their experiments
is infinitely more interesting than any results to which their experiments
lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20100112/2baa5cea/attachment.html>