[mythtv-users] Metadata lookup kills FE

[Robert McNamara]

On Mon, Jan 2, 2012 at 6:16 PM, Michael T. Dean <mtdean at thirdcontact.com>wrote:

> On 01/02/2012 09:09 PM, Kevon wrote:
> > On 01/02/2012 08:46 PM, Robert McNamara wrote:
> >> On Mon, Jan 2, 2012 at 5:31 PM, Kevon<mythtv at reidsresidence.com>
>  wrote:
> >>
> >>> I am experiencing frequent crashes of my FE whenever mythvideo is
> >>> doing a large metadata update. This happens specifically when myth is
> >>> looking up video data at tvdb or tmdb for a large number of videos (so
> >>> far 8 or more seems to be the magic number). I can easily reproduce
> >>> the crash by adding a large number of files to the video directory. I
> >>> get the below error in FE logs after the recent crash:
> >>>
> >>> QObject::connect: Cannot queue arguments of type
> >>> 'QNetworkReply::NetworkError'
> >>> (Make sure 'QNetworkReply::NetworkError' is registered using
> >>> qRegisterMetaType().)
> >>> 2012-01-02 20:04:43.717 Tried to write
> >>> http://www.thetvdb.com/banners/episodes/...., but it appears to be an
> >>> HTML redirect (filesize 0).
> >>> QObject::connect: Cannot queue arguments of type
> >>> 'QNetworkReply::NetworkError'
> >>> (Make sure 'QNetworkReply::NetworkError' is registered using
> >>> qRegisterMetaType().)
> >>> 2012-01-02 20:05:52.896 Tried to write
> >>> http://www.thetvdb.com/banners/episodes/...., but it appears to be an
> >>> HTML redirect (filesize 0).
> >> What Qt?  From the error message, it sounds like the Qt itself is broken
> >> (as metadata download relies on Myth's Download Manager, which is turn
> is
> >> built around all the Qt network functionality... which is what it's
> >> complaining about here).
> >>
> > I am running qt-4.7.0. Is there a way to test of qt is at fault outside
> > myth?
>
> FWIW, the "QObject::connect: Cannot queue arguments of type
> 'QNetworkReply::NetworkError' (Make sure 'QNetworkReply::NetworkError'
> is registered using qRegisterMetaType().)" warning isn't important.
> The code has been changed in unstable/development so it is "proper", but
> whether or not the change is in place has no effect on the metadata
> grabbers' ability to do their jobs, IME.
>
> Mike
>


And in addition to what Mike adds, we'd really need a proper backtrace to
investigate this.  It's possible/probable that it's an image loading
threading issue which is fixed in master, but for which the fix cannot
easily be backported.

Robert
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.mythtv.org/pipermail/mythtv-users/attachments/20120102/b43e970a/attachment.html