Hi Skinah,
i have two “news” on that, one good one bad.
bad first
I gave up comunicating with Reolink support!
Although i pointed them to the very exact protocol element causing the issue, i haven’t received ANY detail.
Nothing whether they tried or even managed to reproduce the issue.
There are plenty of emails they promissed to escalate to designer and to come back, actually without any response, week by week complaining from my side!
Useless to say, that there was no finger pointing or blaming.
good news
I managed to improve/extend the IP camera binding.
It’s now capable of handling multiple IE within one message, i.e. MessageLimit set to higher number than 1
Looks like Reolink camera is stable with this (fingers crossed)
I as well added few further functionalities to the ip camera binding:
-
PET detected (tested)
-
human detected (tested)
-
car detected (not tested - due to lack of car in my garden ;-))
Not sure what API in the thread you are exactly referring to …
I experienced calling the Reolink CGI script to retrieve AI details (pet, human, car detection) at the point in time motion has been detected, in some cases doesn’t work out!
Either the AI algorithm still works (you are to fast calling the script) or the result already passed by (range is about 0 to 6 seconds).
After i got the extended binding working, i even experienced e.g. pet detection event before a motion event was raised (in rare cases even w/o any motion event).
Maybe worth to mention that still Reolink events beings raised do not map 1:1 with what they record (no idea on whether this is limited to RLC-811A).
Hopefully this will be fixed with a newer release.
Haven’t reported this yet due to limited / zero progress on the more serious issue 
The binding extensions i added (similar to initial code) don’t make use of a XML parser.
I don’t feel really comfortable with this, but didn’t want to turn the code upside down.
This would have required more detailed tests with other camera types as well, i fear.
Will continue some tests, this and maybe next week.
Up to now camera seems to be stable, as mentioned above.
For convinience attached some ONVIF test cases (explicitly mentioning MessageLimit being set to 1)!
https://www.onvif.org/wp-content/uploads/2017/02/ONVIF_DeviceIO_Test_Specification_17.01.pdf
https://www.onvif.org/wp-content/uploads/2019/07/ONVIF_Receiver_Test_Specification_19.06.pdf
I wiill show up again end of next week on how to proceed further.
Especially
-
risk on further Reolink interface changes (firmware is in beta stadium) breaking the extension
-
how to do code review
Dependant on above, if and where to apply the changes (trunk and/or backport to recent release branch)?
Best regards,
Dieter