[mythtv] mythtranscode memory leak

[Geoffrey Hausheer]

On Tue, 16 Sep 2003 16:46:56 -0400, "Roy Hooper rhooper-at-toybox.ca
|mythtv/1.0-Allow|" <l7mdhk6lt10t at sneakemail.com> said:
> 
> >The only time this should happen is if you DON't use fifosync (or you
> >don't have a mythtranscode that supports it...see 'mythtranscode --help'
> >to see if it is availiable.  It isn't leaking memory, it is allocating
> >buffers to prevent a deadlock.     I guess the easy answer is, make sure
> >you know what you are doing when using mythtranscode from the command
> >line.
> >If your version of mythtranscdoe does have '--fifosync' then let me know,
> >and I'll look into it.
> >  
> >
> Aha!  I am using an older transcode without --fifosync (I had to back 
> out to 0.11 at some point)... I'll be retrying CVS again soon, at which 
> time I will be all good.
> 
Well, if 'soon' is after Isaac applies the profile-group patch (next day
or so, most likely), you may not be 'all good', since the command line
arguments to mythtranscode have changed, and nuvexport may not work until
it is updated. (It depends on what parameter is passed for '-p' to
mythtranscode).

Hopefully the transition won't be too rough, but it will be harder to
revert back once this change happens, since profiles will be handled
quite differently.  Such is the price of progress, I guess.

.Geoff