- instead of relying on gethostname() to return a resolvable name, grab the IP address from the incoming request so that the address we put in the links in index.html is certain to allow the client to retrieve those documents

- more alphabetization of dependency lists
- add missing gir1.2-glib-2.0
- add other missing deps to gstlal-calibration
- remove libxml2 dependency (why was this added?)

- http server should be started on 0.0.0.0 (= all interfaces), and avahi should advertise it on all attached networks for us, so there's no sense in which any server address other than 0.0.0.0 is correct

- rely on it retrieving the default bottle app instead of passing it

- remove job tag from advertised service name
- use default port number (= randomly assigned)
- add line breaks to organize code

- helpful in identifying a jammed server start-up sequence

- let avahi figure out the host to advertise for us
- use the module default protocol and domain as the default arguments for .add_service()

- need to get port number if randomly assigned

- expose the contents of the GSTLAL_LL_JOB environment variable as one of the advertised service's attributes.

- code-drop to latest stable version from upstream

- add gtk-doc to build requirements
- add .gir file to file list

(well try to, we'll see)

- seem to not be needed, the internal queueing of the adders seems to be
  sufficient

- the modern incarnation of the adder appears to propagate format
  constraints up- and downstream properly so these capsfilter elements are
  no longer required for pipeline format negotiation

- a FIXME says to do so when a problem in the collect pads class has been
  addressed.  that problem has been addressed, but we have also altered our
  use of the GstBaseTransform class so that every input buffer always
  produces an output buffer (just, sometimes, with 0 size), which removes
  the need to have fixed the collect pads class
- this causes the firbank and matrixmixer to run in the same thread when
  the otho-gate-fap feature is turned off, but we generally have so many
  more of these elements than there are CPU cores there's no way the
  additional parallelism could have been used anyway.

- from after every audioresample element

- the audioresampler in gstreamer 1.x no longer produces bad discont flags.
  we can remove all use of lal_nofakedisconts element, but before we do
  this patch adds a test to make sure the bug doesn't reappear

- remove an unused instance attribute to prevent reference count bloat
- fix some typos in documentation

- introduce separate handlers for each signal type to remove need for boolean parameters in call-backs, and to make their exact code paths clear
- correct return value for eos call-back (should be void, not GstFlowReturn)
- move code to attach handlers to appsink elements to its own method so it's not duplicated
- set the appsink elements' max-buffers properties to 1 rather than checking that it hasn't been set to something else in the property list when creating them so that elements constructed elsewhere and passed to .__init__() are configured correctly as well.

- set as minimum required version for gstlal-calibration
- set as minimum required version for gstlal-inspiral

- further efforts to prevent collisions with native adder

- list -devel packages dependencies one per line
- alphabetize dependency lists
- add --enable-gtk-doc to configure command