[mythtv] Very long recordings

David Hampton mythtv at love2code.net
Wed Feb 10 18:30:42 UTC 2021

On Wed, 2021-02-10 at 11:05 +0100, Piotr Oniszczuk wrote:
> > Wiadomość napisana przez Klaas de Waal <klaas.de.waal at gmail.com> w
> > dniu 08.02.2021, o godz. 19:08:
> > 
> > Hi,
> > 
> > Yesterday I updated my backend to the latest master and since then
> > some recordings are extended way beyond their scheduled end time,
> > one of them extended to 21 hours (115GB). Guide data looks OK so I
> > guess it is the recording. Could this be another "std:chrono"
> > induced bug?
> > 
> > Klaas.
> > 
> just want to say „me too”...
> i updated be to current master and i have totally screwed real
> recordings durations.
> EPG, OSD, etc shows as expected, but real recordings are much longer
> (i.e. rec. which should be recorded 8PM-9PM in reality was recording
> 8PM-11:30PM).
> Something is definitely wrong here.... (provably by recent st:chrono
> changes?)

I agree there is a problem, but I can't replicate it here. I have a
bunch of recording scheduled on my dev system to see if I can find one
that records up until the subsequent scheduled recording.

Summarizing to see if there is a pattern. So far we have reports of:

No problem:

DavidH: HDHomeRun (ATSC, CC), Fedora 33, Qt 5.15.2
Klaas (dev) has DVB-C tuner, Fedora 33, Qt 5.15.2
John P: unknown tuners, Fedora 32, unknown Qt
GaryB: external recorded, unknown OS, unknown Qt


Klaas (prod): DVB-C and DVB-C HDHomeRun, Fedora 31 Qt 5.13.2
TomiO: DVB-T2 tuners, unknown OS, unknown Qt
jppoet: unknown
Scott Simpson: HDHomeRun Prime, Unknown OS, Unknown Qt

So far nothing jumps out as being common.

Piotr's bad recording didn't cross midnight local time, and I'm
guessing that it didn't cross midnight UTC.

Has anyone else see a pattern where some recording are correct and some
overlong, or are the overlong recording on the system overlong?


More information about the mythtv-dev mailing list