[mythtv] [Patch] copy and paste bug in channelsettings.cpp

Kenneth Aafløy lists at kenneth.aafloy.net
Sat Sep 25 19:02:28 EDT 2004


On Sunday 26 September 2004 00:14, Kenneth Aafløy wrote:
> On Saturday 25 September 2004 23:50, Isaac Richards wrote:
> > On Saturday 25 September 2004 04:59 pm, John Pullan wrote:
> > > On Sat, 2004-09-25 at 22:29 +0200, Kenneth Aafløy wrote:
> > > > On Saturday 25 September 2004 21:59, John Pullan wrote:
> > > > > I've just noticed that the widget that looks after the hp_code_rate
> > > > > was actually using the fec entry :-/
> > > >
> > > > Hmm, the fec and hp_code_rate is sharing the same data field, so this
> > > > is an incorrect assumption. :p
> > >
> > > Ooops my bad. (Bit confusing though)
> >
> > More than a bit confusing.  That's plain old bad design, there.
>
> Why, they are never used at the same time, since a card can't be both
> dvb-s/c and dvb-t...because then it will have two frontends, which will be
> enumerated in the dev directory as adapterx/frontend0 and
> adapterx/frontend1.

One more thing I want to mention is that the settings api was crippeling me so 
much that I could not make the separate dvb-s/c/t settings screens I was 
hoping to develop. Then the internal usage of data fields would not be 
visible to the user. Also the statement above is a bit false, because card 
should really be replaced with frontend. The MythTV dvb code also does not 
support multiple frontends/demuxes on a single card, which apparently is not 
used in other hardware than real stbs.

The approach of sharing data fields in dvb has been accepted into the linux 
kernel a long time ago, so I guess what your opinion of what is 'bad design' 
is a bit outdated. And you should care not to give these complaints only to 
the dvb code just out of convinience, as I have found several major 
design-flaws in the core-myth code.

Cheers,
Kenneth


More information about the mythtv-dev mailing list