[mythtv] Backend AutoDiscovery

[f-myth-users at media.mit.edu]

By the way, this has been sitting unsent in my editor for a few days
now; perhaps I should at least mention it here instead of starting a
new thread, in case it's relevant:

Subject: Random version-mismatch suggestion

A large number of bugs seem to be caused by accidentally mixing
versions between frontend & backend, leading to a large background
level on the lists of messages asking for help because of them.

I'm wondering if it wouldn't be a good idea to have, right on the
toplevel frontend UI page where it can't be missed, some sort of
thing in a corner of the display in big red letters that says:
  THIS FRONTEND:  version x.y.z
  THIS SLAVE BE:  version x.y.z  [if this is a slave BE]
  MASTER BACKEND: version x.y.w
or something like that, iff the frontend notices that it's running a
different version from the backend.  (And it seems that this would be
a useful thing to have on -some- page -all- the time, even if the
versions appear matched, to make it easy to tell that they are
matched and what all the versions are---but -that- readout should
probably not be on the toplevel UI page, but somewhere where not
everyone who uses the UI will be distracted by it.)

This presumes a number of things:
(a) That "version" is a well-defined concept.  What with the
    differences between point releases and SVN, and between various
    branches of SVN, it's not clear to me that there is a single
    unambiguous way of referring to Myth versions.  I've seen lots
    of confusion on this point on the lists as well.  It would be
    really nice if this info got compiled in on every compilation.
(b) This information is discoverable at all.  It's not clear to me
    that the protocol actually allows a FE to know what version the BE
    is running (except for its protocol version, which isn't the same
    thing).  Perhaps DB-schema version is enough to declare that the
    FE/BE versions are "the same", but I'm not sure if this alone
    actually qualifies as "same version"; do we have counterexamples
    where matching DB-schema and protocol versions still weren't
    enough to guarantee "close enough" versions?  (Ugh, probably
    amongst branches, but perhaps even in mainline combos?)
(c) That various themes are updated to include this information.

[Yes, I know a slave BE w/no FE doesn't have a display to show this,
but at least it takes care of -some- of the combinations...]