> It sounds quite logical.. It wouldn't break anything too. Probably the > best to test this patch would be on lowbitrate streams on satellites > like Hotbird .. ;-) It shouldn't make any difference otherwise. My > antenna which was pointing at Hotbird, went down in a wind a while back, > and i haven't replaced it yet. The case that provokes the problem is where a PID carrying section data contains both small sections, less than a TS packet in size, and large sections, very close to the maximum 4096 bytes. This can arise in DSMCC carousels, in EIT schedule data (EPG information) and anywhere else where there are sections of various sizes. The bitrate of the stream is of no consequence: the data loss caused by the bug will be the same. Since DSMCC and EIT data is broadcast repetitively, there is a good chance that the section lost by this bug will be found on a subsequent repetition. However, this is not guaranteed because in some (many?) cases, the pattern of the sections in the stream will be the same at the next repetition and the bug will strike each time. The situation that prompted me to investigate and propose the patch was a DSMCC carousel in which I found I _never_ received a particular file that I knew was present. Mark
