[mythtv-users] Schedules Direct TBA Listings

[Gary Buhrmaster]

On Thu, Jul 17, 2014 at 6:13 PM, Daryl McDonald <darylangela at gmail.com> wrote:
....
> With regards to " dd--grab-all being the default setting, couldn't the
> individual system upon sensing it's own low power, then revert to previous
> default? Forgive my newbie ignorance, but if possible this seems a simpler
> work around.

While such software is not impossible, and I am not writing the
code, but it seems to me that determining "underpowered" is
actually rather hard to accurately determine.  It is not simply a
slow CPU, "underpowered" can be many factors.  While one
could try a download, and set itself to degrade in future downloads
if it fails (which, as I recall, could leave the system with no
scheduling info at all if it fails in certain steps), a single failure
is not sufficient (because it could be due to an Internet outage,
or an overloaded link due to someone running bittorrent, or /tmp
filling up because someone downloaded the 400TB copy of the
unedited raw Hobbit footage, or /var filling up due to an infinite
number of log entries from ssh attempts to root, or the (shared)
database being overloaded by doing a star join against 100 tables,
or ....)  And then, if someone upgraded their system (or solved
the other issues), one would probably want it to change to
--dd-grab-all automagically.  So you have to add some state, and
some dynamic testing, and some hysteresis to insure that you do
not flip from --dd-grab-all to --dd-grab-old-and-slow-and-obsolete too
quickly.  And, btw, you would not want to be overloading SD doing
a test every hour just to be sure.  All in all, probably more work
than simply doing the right thing and use the already existing
loading capability as Michael suggested.  But, ultimately, it would
be up to the devs to consider your code contribution for
underpowered system determination and fallback, and whether
it meets their expectations for the principal of least astonishment.