Shay Rojansky

Just referencing [this Npgsql refcursor discussion](https://github.com/npgsql/npgsql/issues/438#issuecomment-279542211). tl;dr: Npgsql-specific types aren't actually required for dereferencing refcursors, and single-resultset PostgreSQL functions should really return a table rather than a refcursor.

> Though, one question to @roji, is the NpgsqlBinaryImporter.Write() operation completely supporting the NodaTime without doing any tweaks? Yep, as far as I know there shouldn't be anything special about...

This should be ready for review. I'll continue working on test coverage in the coming days, but the code should be more or less ready.

@GrabYourPitchforks > I also see some stray references to the System.Security.Permissions namespace. Those APIs are all obsolete, so be sure to remove those references. Thanks, removed the last lingering one....

@stephentoub sorry for the big pause, was busy with some other urgent stuff - will now work to merge this ASAP. > That's a lot of code :) I skimmed...

@stephentoub @AaronRobinsonMSFT do you guys have any idea about the build failures? There's various "Managed parameter or return types require runtime marshalling to be enabled" errors reported on other projects,...

@danmoseley yes, I'm planning to merge this for 7.0. I'm currently blocking on flaky CI test failures, and will spend some time in the next couple of days trying to...

The remaining distributed transaction test failures occur only on 32-bit, these are consistent and easy to reproduce locally as well. As discussed offline, I've skipped the MSDTC tests on 32-bit...

Thanks @julielerman :)

Just a big +1 to adding support for properly benchmarking in a concurrent environment - it's super important to be able to see how code behaves under concurrent load (lock...