llrp_manager_example segfaults when there are > 200 targets on Linux about rdmnet

[samkearney]

Oh, good point! I will test your steps on my Windows machine and get back to you.

[samkearney]

@peternewman I was able to reproduce this issue on Windows and I believe I've now fixed it on the develop branch. If you can merge back to your Linux branch and re-test that would be appreciated.

[samkearney]

Thanks Peter. As you know we don't officially support Linux yet so we haven't really done any testing. If you want to investigate this and fix it in your PR, please feel free. Otherwise, we'll be working on Linux support within the next couple of months so we will test and address this issue then.

[peternewman]

Thanks Sam, out of interest, have you tested this particular scenario on your supported OSes? I'm guessing you have? I've not looked at the code too closely, but I'm a bit surprised it would work on one and fail on another OS.

[peternewman]

It seems more subtle than I first thought, I can have e.g. 300 devices total fine, but not 201 of the same code (which is the same manufacturer ID and LLRP type; mine are LLRP only, ETC are RPT Device). I haven't tried more on the manufacturer ID thing yet.

[peternewman]

That's great news. I can confirm I've tested this fix merged into #27 and I now longer get the crash (the most recent test had 201 of my LLRP targets running and one ETC one, but reliably crashed before updating and recompiling and didn't crash afterwards.

Thanks for fixing this @samkearney , I guess that means my Linux work isn't too far off track if it can find bugs in the main codebase on other OSes!