parallel_pg_loader output fails if mfr switch is used and bbb_simple_rec_trigger exists

This bug is present in the 2.3.0 release tarball and will trip up anyone trying to import MARC records by following the standard import instructions.

I think the given commit is on the right track. One thing that concerns me is the history of this code. It looks like we are more or less reverting to what was there before, see here:

http://git.evergreen-ils.org/?p=working/Evergreen.git;a=commitdiff;h=8595342b1153b7d562fb2d3f5a775e80b9c8d030

For more explanation of this change, read this:

http://libmail.georgialibraries.org/pipermail/open-ils-dev/2009-September/005081.html

It doesn't look like the zzz_update_materialized_simple_record_tgr exists anymore, so I don't think that but is valid, but I am stuck trying to understand how the fix ever made sense, and feel like I am missing something. In any case, if we do this fix, we probably want to do the same in pg_loader.pl (if anyone still uses that).

I think I'm missing something, here. In what situations should [parallel_]pg_loader.pl be used to generate COPY statements for anything but 'bre' objects in modern Evergreen?

I ask because the code that could generate 'mfr's is frighteningly out of date, and replaced by in-db ingest. Sorry if I'm misunderstanding ...

I think the fix is probably to warn on, and ignore the effect of, anything but '-or bre' and '-a bre' from the command line, WRT objects we should try to load.

We were doing our first migrations a year ago when this was submitted, and have since dropped all switches other than '-or bre'.

Since the code in question is essentially dead (and, worse, possibly harmful) I've un-pullrequested this and marked it incomplete. Currently proposed plan of action is to trim *pg_loader.pl down to support only 'bre', 'are' and 'sre' objects, warn and ignore others.