David Pacheco

I'm pretty torn about this. I feel I've wasted a lot of time on DWIM-style interfaces, both debugging callers and just reading code written with them. Because this change looks...

Obviously I've been conflicted -- I filed #19. Sorry for the waffling.

How reproducible is this problem for you? Are you (or can you) run with --abort-on-uncaught-exception and share the resulting core file?

I think I understand what you're saying, but without understanding this better, I'm not sure what the most appropriate fix is. (I don't want to blindly add a defensive check...

Thanks. Interesting that they were two different errors.

Thanks for taking a swing at fixing this. I'm not sure I understand the new invariants, though. Before the change, the invariants were that there was only ever one active...

Thanks for doing this. How was this tested?

Thanks for this. Under what conditions are the locations null though?

@AmlingPalantir Given that it's easy to reproduce, could you post the output that came back from Postgres? If it were me, I'd just run with --abort-on-uncaught-exception and look at the...

Given that context, I'm not sure it's sufficient to just set the locations to "-". Wouldn't it make more sense to label the replication kind as "pgbackup" or something to...