Merlin Moncure wrote:
>
> > Similarly, if I have PostGIS or PL/R on the hacker's server, or- heaven
> > forfend- both, is the best way to get at the production server still to use
> > dblink?
>
> dblink allows you to send queries from one server to another in a
> couple of different ways. What the 'client' server has installed is
> irrelevant...the sql is processed by the 'server' server (in your
> example, the production server i think).
>
> now, you could send the data across via a dblink query/view and
> pl/perl process it in your developer box. if you have pl/pgsql
> installed on the production server, I would suggest using that though
> and just invoking a function call across the dblink ;)
Thanks for that. One of the reasons that I am contemplating this is that when I
built the server it wouldn't build PL/Perl since the underlying distro didn't
provide a libperl.so file. Now I could obviously recompile the distro's Perl
sources but that would mean I'd no longer have a common Perl architecture site
wide; I'm considering building a .so on a scratch machine and copying it to the
production server but I'm not confident that I understand every possible
implication.
The other thing that I'm thinking is that it's quite possible that (as
hypothetical examples) PL/Perl, PostGIS and PL/R wouldn't be happy on the same
machine, at which point the only way to merge their functionality in complex
work would be to use a "farm". I hasten to say that I don't anticipate trying
that, at least /this/ year, I'm just trying to think ahead :-)
--
Mark Morgan Lloyd
markMLl .AT. telemetry.co .DOT. uk
[Opinions above are the author's, not those of his employers or colleagues]