[mythtv-users] Schedules Direct JSON service on the back burner?

Wed Oct 15 21:23:23 UTC 2014

On Wed, Oct 15, 2014 at 1:19 PM, R. G. Newbury <newbury at mandamus.org> wrote:
> On 15/10/14 02:16 PM, Thomas Mashos wrote:
>>
>> On Wed, Oct 15, 2014 at 11:04 AM, Joseph Fry <joe at thefrys.com> wrote:
>>>
>>> I have noticed that almost all of the discussion on this list has related
>>> to
>>> the RobertE's XMLTV adaptation to SD's new JSON service... which seems
>>> like
>>> more of a workaround than anything.
>>>
>>> Is anyone actually using the JSON service, is there any effort to move
>>> things in that direction?  It really does seem like the better solution,
>>> though I suspect that we will ultimately need database changes, and
>>> perhaps
>>> some changes to mythtv-setup (or whatever is to replace it) to manage our
>>> lineups.
>>>
>>> While I suspect that I will have to use the XML solution come Nov 1st, I
>>> fear it's simply because adjusting things to support the JSON data isn't
>>> getting the attention it deserves.  I know RobertK has a script for the
>>> JSON
>>> service, but haven't seen any real discussion from folks testing it...
>>> thus
>>> I am not sure I will be comfortable going that route come Nov1.
>
>
>>
>> Moving to the JSON service has a few challenges.
>>
>> 1) It requires a JSON grabber. Now IIRC Robert does have a JSON
>> grabber, but I believe that the idea is to move to using the JSON
>> service via XMLTV (so there is a single codebase for updating guide
>> data in MythTV). This requires more work.
>>
>> 2) Using the JSON service (and thus the JSON grabber) requires more
>> user interaction than using the XML DD replacement service. Most agree
>> that while the JSON grabber is preferred, it's not ideal to require
>> the user to do the amount of work required for the JSON change during
>> a minor upgrade (a more natural time for this change would be the
>> upgrade from 0.27 to 0.28). IIRC, we cannot do a zero impact upgrade
>> for a user and move them to the JSON service, it will require that
>> they setup the new grabber.
>
>
> The JSON service and grabber are still there and working well. The only
> 'user interaction' required is during setup. There are the same number of
> steps as in an original mythtv setup, only some are done in mythtv and some
> with the sd-utility.
> Thereafter, the JSON grabber runs in background without intervention.
>
> <digression>
> Some background for those who missed the prior messages. I am a director of
> Schedules Direct and have learned a lot about how the services actually
> work. in the last few months.
>
> We are actually talking about three different data sources.
>
> 1) We refer to the original and present data feed as DataDirect or TMS-DD.
> In fact, that feed is hosted on computers owned by Tribune, the old parent
> of Tribune Media Services, now owned by Gracenote. Tribune wants their
> machines back!
> This is the service which is going away.
>
> 2) Robert Eden accesses a different data feed, (the 'flat file') parses it
> and loads it onto servers we run. (I think he has to transform it to XML,
> although it may already be XML). This feed is North America only.
> There are two levels of 'fix'.
>
> Fix A:  The RE fix at first instance is to point the mythfilldatabase
> grabber at the IP address of 'our' server (and bypass DNS looking up the
> real address). (There are other methods using RPZ, too)
>
> Fix B   Mythtv devs have already committed patches to the code base, to
> point to the new correct domain. Unfortunately, this fix is only for .28-pre
> (aka master) and fixes-.27-4. This was inserted into the code base 2 days
> ago. Fixes-27.3 was bumped to 27-4 earlier today.
>
> The requires compiling your own programs, or re-installing mythtv with the
> correct version. For Fedora (etc) users atrpms is still on .26 while
> rpmfusion is on 27.3. I don't know where to look for Ubuntu/mythbuntu etc.
>
> If you do not intend to compile your own, and do not wish to upgrade to
> something with the fix then you will have to use Fix A. (I strongly suspect
> that pre-packaged 27.4 versions will be available before the end of the
> month for those upgrading.)
>
>
> 3)  Robert Kulagowski's grabber accesses a completely different feed which
> TMS had called ON and which Gracenote calls On-Entertainment. This feed
> covers 30 odd countries already and has a much wider available breadth of
> meta-data, most of which has not been built into or available in myth. One
> thing which is presently available is the presence of unique hard-coded
> series and episode numbers for tv programs.
>
> </digression>
>
> I have been testing RK's JSON grabber since late April, and using it for
> daily test purposes since mid-July. I have been using it at home on the main
> box since early September. NO PROBLEMS.
>
> Yes,there are a number of steps during the install. You did them when you
> set up your box. You do not have to touch anything to do with the database.
> Going back and forth from mythtv-setup to the sd-utility seems confusing,
> but the sd-utility just replaces parts of the mythtv-setups videosource
> setup stage.
>
> The things which seem to have evidenced the most confusion involve:
> 1) deleting all video cards, and
> 2) the deletion of video source data.
> You do not *have to* delete your video cards. But they *will* become
> unlinked from your present video sources.
>
> You cannot have both old and new setups active at the same time, so your old
> video sources should be deleted. Your new video source is no longer TMS-DD.
> The lineups have slightly different names but lineups are a
> choose-from-the-list, enter-the-number' proposition in the utility. If you
> choose entirely new names for your sources, your old sources can be
> retained, but as dormant rows in the tables.
>
> One possible point of confusion regarding Antenna/OTA. A scan is required
> because we have no way of knowing which tv channels you can actually
> receive. And we can set up custom lineups if you can get something which
> others nearby cannot receive.
>
> Another point of confusion, is that with the JSON setup, when you select a
> cable type lineup, ALL of the channels in that lineup are included. You have
> to use the channel editor to set the ones you do not want, to not be
> 'visible'. That includes OTA channels which the scan picked up but which you
> do not want to watch.
>
> Once set up, the grabber substitutes for mythfilldatabase completely, and
> runs every day without problem.
>
> The JSON feed is not limited to North America, but covers basically all of
> the Caribbean, most of South America, and a number of countries in Europe
> including Great Britain.
>
>
>
> --
>              R. Geoffrey Newbury
>
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://www.mythtv.org/mailman/listinfo/mythtv-users
> http://wiki.mythtv.org/Mailing_List_etiquette
> MythTV Forums: https://forum.mythtv.org

While I agree with most of this, my point was that users will update
from packages on a fixes branch and expect not a lot is required of
them to get what are essentially bug fixes (eg. no new features).
While the move to the JSON grabber is one that is done a single time,
it does require the user to go forth and actually do that setup.
Further, somehow we must notify the user to go and do this setup.
While some users may be confused about the amount of work that needs
to be done, let me clarify that I am not confused about the minor
setup work users must do, I just disagree with making them do that
work at any point other than a major MythTV version upgrade (eg. 0.27
-> 0.28).

The Mythbuntu team provides daily builds of the fixes branches for
various releases of Mythbuntu (see http://www.mythbuntu.org/repos ).
The fix was released to MythTV 0.27 today and Mythbuntu had builds
available for users within hours of that fix. That means that 90%+
(not counting those who build from source) of the users have the fix
available to them (although it appears only about 20% take advantage
of those builds). We've also done a blog post to try and highlight the
issue for users not using those daily fixes builds.

While I agree (as I think most people do) that the best way is to move
forward is using the JSON service, I also believe a zero impact fix is
needed until a major upgrade. Simply because upgrading major versions
of MythTV is more of a user involved process. MythTV fixes can be
received by a simple apt-get upgrade, however a major MythTV version
upgrade requires a new repository to be enabled for most users.

-- 
Thanks,

Thomas Mashos