[Kstars-devel] HTMesh Perl module?

[James Bowlin]

On Mon, Jan 12, 2015 at 10:43 PM, Akarsh Simha said:
> > This was back in the KDE-3.x days.  It used the dcop
> > interface to draw HTMesh elements in KStars which was pretty
> > cool.  AFAIK, none of that interface survived the switch to
> > dbus.  But that is independent of the HTMesh.pm module which
> > was just a wrapper to let Perl programs use the HTMesh
> > library.
> 
> I thought we had some of that code commented out, which was
> later removed -- but git knows about it, so we can resurrect
> the trixel plotting stuff any time we need to.

Not unless the interface inside of KStars is reimplemented using
dbus.  I'm pretty sure the decop interface no longer exists so it
is not just a matter of uncommenting KStars code.  IMO letting
people draw on the KStars sky via an easy external interface is a
very cool feature. I used it a lot for debugging but this is
way OT here.

> KStars still uses HTMesh. It seems like the C++ version of the
> library is unmaintained. A search for Hierarchical Triangular
> Mesh library leads me to the SDSS website, where I can only
> find C# versions of the library. So I'm not sure if there are
> newer C++ versions.

I agree.  I don't think anyone else is maintaining the C++
version.
 
> We seem to ship a HTMesh library with KStars. This builds fine
> and works fine. This was stripped down quite a bit in the
> recent past because KStars does not use all the functionality
> of HTMesh. I don't know whether stripping down the HTMesh.xs
> perl module is a better solution, or whether reintroducing the
> removed methods into the HTMesh library shipped with KStars is
> a better solution. 

That is almost exactly what I said to Jasem in response to a
private email from him:

   If the problems compiling the binary part of HTMesh.pm are due
   to "missing functions" then the solution depends on which
   functions are missing.  If they are superfluous to the
   functionality you need then just remove those unneeded parts
   from HTMesh.xs (which is the source code for HTMesh.pm).  If
   the missing functions are required for the functionality you
   need then someone is going to have to restore that
   functionality to the htmesh library.  

    If you show me the error messages I can try to help advise
    you what to do next.

If the problem can be fixed by trimming down HTMesh.xs then I
think this is by far the easiest solution.  Why resurrect code
we don't need and won't exercise?  Sure, maybe someday someone
might want it but I don't think it makes sense for us to do extra
work for that possible (but not likely) eventuality.


Peace, James

              Science is the belief in the ignorance of experts.
                                             -- Richard Feynman