Artem Vorotnikov

@tomusdrw Is it correct that `unsent_pooled_hashes` is an unnecessary addition in general and the better approach would be to compute the difference between `last_sent_hashes` and transaction queue?

> By "Option B", do you mean option 2? Yes

What if a block timestamp is forged but still fits `i32`? This code can't catch this in any way.

Can you test with the latest master?

@davidbarsky @mitsuhiko can you please merge this pull request so it can be a base for others?

This will cause additional allocation on each peer, what is the capability in question?

Yes, it looks like specific to Quorum, but we don't support their protocol anyway?

I support changing the default to match the rest of GNOME.

Deduplication of dependencies when used downstream.

@adetaylor done