Zeeshan Ali Khan

Hi @jhartzell42 :smile: I'm very glad that you didn't just give up after our disagreements. :smile: > So I continued working on solving this on a fork of `zvariant`, and...

In GitLab by @jhartzell42 on Jan 19, 2022, 19:58 Yep, that's one of the rough edges for sure. So the current code definitely does the wrong thing (although the array...

In GitLab by @jhartzell42 on Jan 19, 2022, 20:08 I plan on adding documentation, TODO lists, etc. to this repo over the next two days, in case you'd like to...

Thanks for explaining. As you said yourself, requiring `SerializerPolicy` isn't very different from requiring `Type` so if you require it, IMO (you can disagree of course) it's not really much...

In GitLab by @jhartzell42 on Jan 19, 2022, 20:53 How often are empty arrays used? I would guess that they are not strictly needed that often. I plan on changing...

In GitLab by @jhartzell42 on Jan 19, 2022, 21:14 Another advantage is that you could write one `SerializerPolicy` for the entire application, and combine your application's `SerializerPolicy`s with an existing...

> How often are empty arrays used? I would guess that they are not strictly needed that often. If a set is empty, it's empty. It's not a matter of...

> If you do send an empty array, it will send the wrong signature in that case (probably `av`, but whatever makes the most sense), unless you specify `SerializerPolicy` to...

In GitLab by @jhartzell42 on Jan 19, 2022, 21:45 Very good point! And I guess I'll start working on my version of a derive macro to make this easier. Sigh....

> Given that, I suppose it would not be a great improvement from your point of view, although still worth it from mine. Would you be willing to accept MRs...