Re: [Live-devel] bug in RTCPInstance::processIncomingReport (RTCP.cpp)

Спасибо! On Thu, 17 Nov 2016 02:46:56 -0800 Ross Finlayson wrote Frederik, Thanks for the report. Your test program helped a lot. There was a bug in the handling of RTCP “APP” (sub)packets. I’ve just installed a new version (2016.11.17) of the “LIVE555 Stream

Re: [Live-devel] bug in RTCPInstance::processIncomingReport (RTCP.cpp)

Ross, I'm happy to hear you've fixed it so quickly. I was about to check if the buffer's packets were valid using gstreamer gst-rtcp-buffer-validate-data-reduced: https://gstreamer.freedesktop.org/data/doc/gstreamer/head/gst-plugins-base-libs/html/gst-plugins-base-libs-gstrtcpbuffer.html#gst-

Re: [Live-devel] bug in RTCPInstance::processIncomingReport (RTCP.cpp)

Frederik, Thanks for the report. Your test program helped a lot. There was a bug in the handling of RTCP “APP” (sub)packets. I’ve just installed a new version (2016.11.17) of the “LIVE555 Streaming Media” software that should fix this. Ross Finlayson Live Networks, Inc. http://www.live555.c

Re: [Live-devel] bug in RTCPInstance::processIncomingReport (RTCP.cpp)

Hi, I figured out that the RTCP packet was not the first packet (*pkt) in the buffer meaning that we may have a compound RTCP packet with individual subpackets. I've copied the entire 1456 bytes of fInBuf and now I can simulate the issue offline, the test code you can find in attachment. C

Re: [Live-devel] bug in RTCPInstance::processIncomingReport (RTCP.cpp)

Hi Ross, unfortunately I could not find any packet that allowed me to reproduce the issue (feeding hard coded data to processIncomingReport with info I'd collected from memory right before the crash). What looks like a quick fix to me is the following: there is a code line in RTCP.cpp::proce