<div dir="ltr">On Thu, Aug 15, 2013 at 12:22 PM, watch_the_rain <span dir="ltr"><<a href="mailto:watch_the_rain@163.com" target="_blank">watch_the_rain@163.com</a>></span> wrote:<br><div class="gmail_extra"><div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Oh,no,the data are saved in a file on disk for testing purposes...that's why I'm puzzled, the running environmens are all the same.<br>
</blockquote><div><br></div><div>I am not sure about that, since there is no mention on how the stream was received delivered etc....<br><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div class="HOEnZb"><div class="h5">At 2013-08-15 18:03:46,"Andrey Utkin" <<a href="mailto:andrey.krieger.utkin@gmail.com">andrey.krieger.utkin@gmail.com</a>> wrote:<br>
>2013/8/15 watch_the_rain <<a href="mailto:watch_the_rain@163.com">watch_the_rain@163.com</a>>:<br>
>> Thanks for your answer.My boss have changed a data packet to test.<br>
>> But a software my company used to test the TS stream can parse out all the<br>
>> events from the discontinuous packets,why can't the libdvbpsi do this?<br>
><br>
>Because it is not OK to require valid outputs from invalid inputs.<br>
>There are too many things to check, which you said nothing about.<br>
>For example, it may be your program which uses libdvbpsi reads UDP<br>
>(right?) input stream with sufficiently large intervals, resulting in<br>
>lost datagrams. It happens if you do some CPU-intensive processing in<br>
>the same thread that reads from UDP socket. The higher is bitrate, the<br>
>more is probablity to have such problem.<br>
><br>
>--<br>
>Andrey Utkin<br>
<br>
</div></div></blockquote></div>Kind regards,<br><br></div><div class="gmail_extra">Jean-Paul Saman<br></div></div>