[mythtv] Debugging DeviceReadBuffer.cpp

[Thomas Börkel]

HI!

On 31.05.2013 19:52, Daniel Thor Kristjansson wrote:
> From the looks of the patch you don't print anything for
> the (ret > 0) case. It will be fairly rare for 2.5 seconds

I added this now for testing and got lots of debug messages. So, the
patch works and the debug messages work in principle.

Also, I printed out the actual poll time according to MythTimer and it
is 20-30ms.

So, if I understand the code correctly, the first poll should be with a
10ms timeout and the 2nd try with 2500ms.

I think this is another proof that poll() does not care about the
timeout, because otherwise I should get a retry every time.

> to elapse with no data coming from the device. You generally
> need to unplug the aerial to get to those cases.

I don't exclude that my SAT equipment may have a problem every now and
then, because it happens mostly on 1 or 2 multiplexes, but I think that
MythTV should better handle this and it seems the code is written with
that in mind. It's just that it does not work, because it does not retry.

Regards,
Thomas