trig2disk reads type_triglist messages from the hypo_ring (e.g. those produced
          by arc2trig and by carltrig) and writes either sac or ah triggered
          data files to disk.


The ah data is in xdr format which is transparent between sun/solaris and 
intel/NT so long as the ioroutin.c interface is used.

Incoming data are determined from the wf->datatype and outgoing byte order
are determined by the intel or sparc defines in the make.  NT stupidity
is handled based on whether solaris or winnt is defined in the make
(in some cases, lack of winnt define assumes solaris).  Unfortunately,
there are some dual make defines that should be cleaned up; define
both winnt and win32; define both intel and i386.  The ahputaway
in this directory handles byte swapping of incoming data, previous
versions did not.

The NT port of the xdr routines was accomplished by modifying the OpenBSD
code from the ftp site at umich.  Thus modifications must include this
source should be careful not to remove the copyright statements and can't
be sold. Cavalier commenting of pieces of this source that weren't needed
to write local xdr is ubiquitious.  Thus it is not rpc compatible.  When
ftp'ing from NT to solaris, make sure you transfer as ascii to avoid the
goofy ^M characters corrupting each line (else run dos2unix on the binary
transfered file).

sacputaway has been modified from trig2disk. There are 2 byte order issues:
 1) what arch are the data from and what arch are we running on
 2) what arch are we running on and what arch do we want to read on
The first is taken care of by the original sacputaway with INTEL or SPARC
makefile defines.  The second is taken care of by a mod that uses a 
SWAPIT makefile define.  Blindly swaps before writing to disk if SWAPIT is defined.

EW v3.3 is required, specifically with respect to the author field in
the type_triglist messages (e.g. in parse_trig, carltrig, arc2trig).

Will get restarted if restartMe is in the .desc

To do:
  Adding a list of stations to the trigger message is kludgy.
  Not much error checking is done, particularly in the putaway routines.
  How about some documentation, eh.
  Add author subdirectory to data directory e.g. ah/author/datatime/scn.ah
  The p in memphis.pscn isn't used yet.
  The output data directory is only two digit year
  Some files in the run_test get corrupt headers and I'm not sure why.
