- add tail optimization to .add_ratebin() to try to improve trigger generator performance especially with large numbers of disjoint trigger segments.
- NOTE:  the correct fix is for injection jobs to not record any trigger statistics at all, but that would be t0o invasive a change at this late date.  something for O3

- the newly resurrected stock python plugin loader induces segfaults in gstlal if the two are installed together, and there is no work-around.  we will be transitioning to the stock loader for O3 but for the time-being we need to add the conflict to the package data to prevent breakage.

gstlal_compute_strain:  added option to remove DC component from residual and control channels before filtering.

the user to specify columns will will be used for gps times (and thus should
have no commas) and columns will should be displayed in scientific notation
(currently, that is only the FAR and FAP columns in the candidate table).
Modified gstlal_inspiral_summary_page to take advantage of these new kwargs

lal_resample:  found downsampling bug that counts samples incorrectly when the first timestanp is not a multiple of the longer sampling period

datasource.py: Added rate kwarg to mksegmentsrcgate
multirate_datasource.py: Pass max rate to mksegmentsrcgate

added gstlal_etg_whitener_check exec to check output from whitener, changed Makefile.am to reflect this

- expand tests of trigger_rate module to cover .coalesce() method

gstlal_etg + multichannel_datasource.py: fixed issues reading in channel INI files, added option --channel-include to specify specific channels to read in from INI files

idq_multirate_datasource.py: commented out initial low-pass filter since it is not clear that we get the intended behavior for output triggers

gstlal_inspiral_injection_snr: made number of parallel processes a command line argument. Default = 1

gstlal_inspiral_pipe now uses reference_psd.cache for gstlal_median_of_psds