- The trigger processing/writing function has an option to flush the
  trigger buffer after writing -- it is on by default
- Since triggers which occur after the file span still remained in the
  buffer, they were dropped mistakenly
    - This would only happen at file boundaries. It was noticed because
      the triggergen had been rewritten to output much larger buffers of
      triggers.
- The fix is to make the default flush=False
- Ideally, this should just be removed

not really used for anything yet, and there will be corner cases thees
don't handle properly, but I keep writing this code over and over in test
scripts, so it'll be handy to have them around.

at some point I thought we'd worked out that it didn't matter, but I can't
reproduce that calculation and when I gave it a try the PDFs came out
differently, so let's randomize over psi.

- move a bunch of attribute, method, and dictionary look-ups out of the
  intstrument combination "interpolators"

Kipp Cannon authored 10 years ago and Kipp Cannon committed 10 years ago

replace with ncx2

- replaces use of Gaussian kernel to approximate noise effects with proper
  distrubtion, but will make code quite a bit slower.  we'll worry about
  speeding it up later, better to know it's right.

lloidparts.py: update mkLLOIDmulti() to have a default h(t) gate threshold of infinity instead of None

gstlal_inspiral: update gstlal inspiral to have a default h(t) gate threshold of infinity instead of None

- also, don't forget to regenerate the combined rates after updating
  zero-lag counts