* Reflect the updated location of the python package (python/client)

 * The cur_block field was being updated before the data, allowing a client to see a reference to an incomplete block.

 * Automatically dumps data messages as it comes in.

 * As an aid for interactively looking at the data dump buffers in a python repl
 * not set to be installed yet

 * This will compare two memory segments data where ever they have overlapping time slices.

Work on issue 11, updating daqd_stream_probe.

Closes #11

See merge request jonathan-hanks/daqd_stream!12

This, was a source of difference between runs of daqd_stream_server even when there were no real changes.

Closes #11

 * Adding new modes to dump various bits of state
 * Moved daqd_stream_probe to its own application directory
 * daqd_stream_probe can work off of copies of the shmem to read dumps from the past

Second and 16Hz data reads now use their own status structures.

See merge request jonathan-hanks/daqd_stream!11

Second merge for split client/server

See merge request jonathan-hanks/daqd_stream!10

Restructuring code to split the client/server dependencies.

See merge request jonathan-hanks/daqd_stream!9

 * This is to allow second data to return more than 1 status value (1 per 16Hz segment)

 * Move files around to install header files for the plugins and server libs
 * Due to CMake constraints have to install the daqd_common library
 * Renamed daqd_stream_cxx to daqd_stream_client & daqd_stream_server

 * Updating the doc generation

 * Client code no longer requires curl, framecpp, ...
 * Added a 'common' helper library to hold shared coded between server/client.
 * Tests pass
 * Still need to work on the install/packaging.

Update second fetching code to return a proper gps time.

Closes #8

See merge request jonathan-hanks/daqd_stream!8

closes #8

 * Previously in a single source context gps 0 was returned for the first entry, not the actual gps time.

 Work on issue 8

 * Moved the logic for daqd_stream_probe into the same directory as the CMakeLists.txt file
   * This fixes an issue when building src dir == build dir

 * No longer need to create and sort a global channel list for each request.
 * The code works with the individual source lists without a global merge.

 * Move channel search code to a separate function as a precursor to other work.

 * add daqd_stream_unit_tests as a ctest defined test
 * updated daqd_stream_unit_tests to run, removing a partial multi-source test
   that is handled in specific multi stream tests.

 * change the client_base api to have a core data retrieval that does not reset the status or do byte swapping.
 * do all status setup and byteswapping at the 'top' level apis, not common/core data retrieval apis.

 * Update the second reading test to alternate byte order and verify output
 * Update the 16th reading test to alternate byte order and verify output

 * Multi stream second data was only doing endian changes on the 1st 16th second block.

 * Changes in a sample/test application.

 * A loop iteration variable was not being updated.  Which was resulting in data being placed in the wrong location for second requests.

 * dcu_identifier had a faulty operator<
 * Updated test programs
   * simple_frame_generator.cc can now take external time from the cds /dev/gps device on request.
   * sample_client.cc expanded to do more testing