[mythtv] MythTV life-cycle support intentions

Lennart Sorensen lsorense at csclub.uwaterloo.ca
Thu Aug 23 16:38:12 UTC 2012


On Thu, Aug 23, 2012 at 11:56:15AM -0400, Anthony Zawacki wrote:
> Perhaps my background colors my thinking (I am a developer,) but
> rolling back is not a normal course of action.  The idea that
> recordings that were made by the "new" system are not listed as
> recorded by the "old" system makes perfect sense.  You rolled back
> the transaction, now the system doesn't know anything about the
> recordings.  I don't think it's unnatural, and I don't think it
> would make sense to try to support it. it should be a rare occasion,
> and you should be knowing what you are getting into before you
> upgrade / rollback, and you should be prepared for the results.

Of course not supporting that is what makes a lot of users not want to
test newer versions and then bugs are found right after a new release
comes out because it wasn't tested in that scenario.

Not sure how one would support it, but it would certainly increase the
potential pool of testers.

> I agree with your sentiment, that if it was easy, then more people
> would test the pre-GA code.  Most of the people that are currently
> testing pre-GA code probably fall into two camps:
> 
> 1)  Those that have a secondary MythTV system, where one MythTV is
> in "production" and the other MythTV is a test system.
> 
> 2)  Those that do not have families that rely on the MythTV system
> for their TV watching enjoyment.
> 
> I don't think Brian (or myself for that matter) falls into either of
> these camps.  When running pre-GA code, you have to be willing to
> accept glitches as bugs are worked out.  If MythTV fails to record
> the favorite show of your family, your so-called WAF will fall
> dramatically.  Yet that is a real risk in Pre-GA code.

Yeah too risky for me to try.  I sometimes think I need a second mythtv
box for testing, but that would require a lot of expensive tuner hardware.

-- 
Len Sorensen


More information about the mythtv-dev mailing list