Brandon Ording

Given that one scenario around public interfaces is allowing you to provide your own implementation of the interface, I'd vote for including them as a potential break.

> the version in LibGit2Sharp.NativeBinaries v 2.0.314. that was referenced by LibGit2Sharp 0.26.2 FYI, LibGit2Sharp 0.26.2 references LibGit2Sharp.NativeBinaries 2.0.306, not 2.0.314. The libgit2 hash in 2.0.306 is 106a5f27586504ea371528191f0ea3aac2ad432b.

>Publishing and consuming from a queue that's being redeclared using a different type is very likely to introduce a race condition between things that never coordinate. I don't think it...

> Looks like the second channel isn't detecting the deletion of the classic queue, either quickly enough or there is a bug causing it not to detect it at all....

> So if you concurrently delete the queue and re-declare it with a different set of arguments, it can trip up an ongoing operation on another channel. How does that...

> Channels do not wait for queues to "apply" their operations. Maybe in some cases they could. I'm sorry if I'm being dense, but I still don't understand how this...

Hi @robbaman, The learning persistence is not intended for the scenario you're describing. It is going to be different than what you are using in production, so it's not going...

Looking at the LicenseManager class, I'm not sure it we'd want to directly expose everything in there, but the suggestion to have a way to provide license compliance information presented...

I think the larger issue here is why you're wanting to exclude assemblies in the first place. The current design is intended for you to not need to do anything,...

Doing some quick research, I do see some historical discussion on this topic, and I see the now-vestigial `Connection.HandleDomainUnload` method. I agree that it's odd for this library to directly...