Modifying ligolw_segment_insert_dqsegdb to save xml file in debug mode; Changing InsertMultipleDQXMLThreaded call to incorporate process_params table data from xml, whoo

Defaulting summary_comment='-' in parseKnown to match old segdb results.  We don't have segment_summary comments any more in the new DB, we have a flag name comment and a flag_version comment, but the server returns just a single comment associated with the flag version.

Removing old comment in ligolw_segment_insert, fixing up QueryForDQSEGDBInsert so all imports and definitions are at the top.

Removing old, erroneous comments in ligolw_segment_insert.  These comments indicated that appends still needed to be implemented, but appends and inserts are handled similarly through the python API being used in the client side calls, with a failed insert falling through to an append automatically.  The correct time-segment checking is already done earlier in the code.

attempting to make ligolw_segment_query compatible with s6 call when you ask for a flag that doesn't exist

attempting to make ligolw_segments_from_cats compatible with s6 call when you ask for a flag that doesn't exist

Failed attempt to make library calls work even when no proxy exists for a user, need to rewrite proxy cert handling to avoid this problem entirely.

Committing untested changes to put and patch methods that use SSL certs.  This only required a very small change to call the new HTTPS handler that is overwritten by the new classes.

First successful attempt at adding https cert or proxy cert authentication to GET method.  This solution is *not* useful for something like a web server that my wish to switch the cert file used to authenticate to the server in between uri calls.  Otherwise, it seems to work.

(cherry picked from commit ca1177e4)

Fixing one part of bash script in publish command and adding for loop to check for existance of log files.  This was designed to prevent the condor jobs from quitting before the forked jobs were finished, but it does not seem to be working correctly yet.

Attempting to fix arbitrary publish generator to run multiple jobs per core.  There is a bug in the use of offset, but it appears to affect differnt cores differently.

Attempting to add repeated background runs on a single core to allow for better parallel publish tests to the ligol arbitrary publish generator code.