BBBSnowball

> I would also like it extended to include QoS 2, it feels incomplete otherwise. That may be useful and PUBCOMP indeed feels like the PUBACK for QoS 2 (for...

The simple change would be this (and I think @ralight was also going for that): 1. Skip the call to `_send_pubcomp` in `_handle_pubrel` if `auto_ack` is disabled (same as for...

> so my initial patch, even though it worked in practice, is wrong. I would argue that your patch is good. There is nothing that says that the library has...

> People should just appreciate that the small fix I did was really simple Oh, we do. I like it very much and want it to go in. That's why...

With the current implementation, puback says that the message has reached an intermediate point between the server and the application. If we say that the paho library is the receiver,...

> At worst, for now, a few mebibytes are lost. I don't think many will mourn those mebibytes at the moment. Yes, I agree. We are writing our own partition...

I like the idea, done in https://github.com/BBBSnowball/flake-compat/commit/d1a6788892881f2d15836d0ac1849d844d7b01de However, this would cause inconsistent behavior for `"relative/path"` vs. `".hidden/path"`, right? Should we rather test with `if builtins.substring 0 1 info.path != "/"`?...

> isn’t there a story about that being the initial reason why paths starting with "." were hidden ? grin I didn't know that. Something like "let's hide `.` and...

Thanks :)

> If I understand correctly, this now allows for completely arbitrary paths, doesn't it? Yes. The old code allows arbitrary absolute paths. This PR adds support for arbitrary relative paths....