search

jitsi / libjitsi

Advanced Java media library for secure real-time audio/video communication.
Apache License 2.0
628 stars 281 forks source link

Debug logging of packet mishandling round 2 #483

Closed gpolitis closed 5 years ago

gpolitis commented 5 years ago

partially reverts 96ffe607. 96ffe607 adds some logging around rtp encodings matching that has some negative side effects. The logging logic is best suited for the caller so it has been/will be moved there.

gpolitis commented 5 years ago

isn't even 1 line per packet going to be a lot?

yes! under certain circumstances we were doing that before (by mistake), and this PR attempts to fix that

bbaldino commented 5 years ago

isn't even 1 line per packet going to be a lot?

yes! under certain circumstances we were doing that before (by mistake), and this PR attempts to fix that

Oh, I thought all this did was, in some cases, don't format the logging of the entire packet (but it would still take up at least one line)?

gpolitis commented 5 years ago

isn't even 1 line per packet going to be a lot?

yes! under certain circumstances we were doing that before (by mistake), and this PR attempts to fix that

Oh, I thought all this did was, in some cases, don't format the logging of the entire packet (but it would still take up at least one line)?

you're right, it does that (i think you're referring to the format parameter that i added to the toHexString method, which avoids adding spaces and newlines for the hex dump), but it also removes the warnings from the mediastreamtrackdesc when they fail to match a packet to an encoding (which is what was filling our disks). now that is done in the adaptivetrackprojection after we've looped over all mediastreamtrackdesc instances.

gpolitis commented 5 years ago

jenkins, test this please

gpolitis commented 5 years ago

jenkins, test this please