compute chisq for the first sample in a trigger window, itacac will now try to
save enough samples to form a coincidence with something in the next window.
The bookkeeping is 'done', though it's probably got bugs. This version of
itacac will also probably still not compile. It's got some coincidence stuff in
it, but the coincidence stuff isn't complete

- add hacks (some of which might prove to be permanent) to filter out
  sub-threshold triggers, preventing them from contaminating the ranking
  statistic PDFs

gstlal_peak_state structs. Not sure if this plumbing will be used, but
committing it so that we wont lose it. Note that itacac will not work at all in
this commit, probably wont even compile

gstlal_ll_feature_extractor_pipe + online fx makefile: add options to disable kafka/zookeeper jobs and switch between condor universe

- a file was added but not included in the packaging rules so it was
  missing from the tar ball.  this seems to fix

- this version does not yet exist, but will have to exist because we need
  the .template_id attribute of the SnglInspiral object.
- also, the migration of ligolw_segments from pylal to glue borked the
  packaging and the problem went resolved, now the only solution will
  require a re-release of glue.
- so ... we require 1.59.3 even though it doesn't exist

- this is a lie, we require 3.28, but it's not available and we can limp
  along with work-arounds until it is.  3.14 is what's in the current
  Debian LDG reference OS, so that's what we'll say we need.

- replace dependency on glue-ligolw-tools with python-ligo-lw now that the
  latter is in production

gstlal_inspiral_pipe: add svd files to the create prior dist stats and fix some remaining ilwdchar compats

lines from header file. Added logic using a local copy of struct
gstlal_peak_state with includes a new boolean called no_peaks_past_threshold,
meant to be a temporary hack to inform python layer of times when a trigger
could exist for an IFO but none do

- teach configure scripts to check the version number of glue and
  ligo-segments, not just that they are installed.  the checks only produce
  warnings at this time.

- add explicit dependence on the python-glue-ligolw-tools package

- switch from glue-segments to ligo-segments package

- elements must be installed using python3, but rest of this package
  requires python2 so we can't switch configure.ac over to it yet.  put a
  temporary hack into the gst/python/Makefile.am script to make it use the
  python3 interpreter during module install and byte-compile