Andreas Oberritter wrote: >On Fri, 2005-10-14 at 18:29 +0400, Manu Abraham wrote: > > >>>>If the PMT contains no CA descriptor on programme level and common CA >>>>descriptors on stream level, the stream level common CA descriptors are >>>>now moved to programme level. >>>> >>>> >>>> >>>> >>>This is wrong, because different ES PIDs can use different ECM PIDs. >>>Moving the CA descriptors to the program level destroys their >>>association to an ES PID. >>> >>> >>> >>> >>> >>So what would you suggest.. ? >> >> > >Your quotes made clear to me that this is a hardware limitation for >which this looks like an acceptable workaround, provided that the >"merging" code is not meant to be used with any other (i.e. not broken) >model. > > > >So I just suggest not to buy this hardware ;-) > > Yeah, Some of the very first models had that limitation. For these cards, yes it is a workaround, but not for the later ones. Decryption at stream level (That is the appropriate term ?) fails for all hardware, (since that is probably not decryption at stream level itself what we are looking at, for that mentioned provider) with that stream, unless handled that way. For these cards it is *not* a workaround If you read EN50221 carefully, you will find that this is not a workaround ! So i need to strike a balance where it works for all cards supported by the driver as expected, not just one bunch of cards. Something that works with everything and something that works as expected. That was what i was looking for. (The driver supports a wide variety of them, Technisat Skystar I, Pinnacle Sat CI, VVMER, Tomato, Twinhan, Hercules, some few that came to my mind) So one cannot say which one is which, even if the vendor provide card models which have that limitation. So with that, i think it should be working with all cards. I do remeber vaguely, somebody once said applications do use it that way. Maybe i need to dig my mails on that. It's been a while now, sorting it out can be a big pain. Regards, Manu
