Use KnowHrp instead of Network

[tcharding]

We have a bunch of functions that take Network when what they really want is something that can be converted to a KnownHrp.

Make KnownHrp public and accept impl Into<KnownHrp>.

[coveralls]

Pull Request Test Coverage Report for Build 7621632788

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.004%) to 84.262%

---

Totals
Change from base Build 7616194895:	0.004%
Covered Lines:	19221
Relevant Lines:	22811

---

💛 - Coveralls

[Kixunil]

ACK 20a5f1f

[Kixunil]

IMO these tests could've stayed but whatever...

[tcharding]

Oh, you are right. I did the change before I used impl Into.

[Kixunil]

Note that I've just realized we'll have to make Hrp public at some point too. But I'm still not convinced it belongs to bech32 and there's likely a cleaner way. But that's for far future, units and primitives are more important now.

[apoelstra]

But I'm still not convinced it belongs to bech32 and there's likely a cleaner way.

HRPs are an integral part of bech32. Where else could they belong to?

[apoelstra]

ACK 20a5f1f

[Kixunil]

Where else could they belong to?

Something like bech32-hrp which can get stabilized sooner than the rest for instance. :)

[TheBlueMatt]

As discussed rather extensively on #2225 I don't think we should do this. Users being forced to move between two different Network enums that have the same purpose isn't a cleaner API, its just more fluff and crap to go have to read in the docs. If we add a new Network, we'll clearly want a KnownHrp for it, and vice versa. Until that changes, I don't see a reason to have two separate enums here.

[Kixunil]

This has nothing to do with #2225 but funnily enough it adds argument for application-specific enums.

KnownHrp already doesn't exist for signet which broke roundtrip. I guess this actually proves my point that Network is not universally implemented. The idea is dead, we might as well rename Network to CoreNetwork to indicate the point.

[TheBlueMatt]

That, rather, implies KnownHrp should be defined for signet, and shouldn't match existing ones. That ship may have sailed by now, but its ultimately a bug in the spec for signet, and not something general that implies Network can't be general in most cases :).

[Kixunil]

Yes, it's a bug in signet but we're already screwed and have to live with what we have. The other possibilities are remove Address support, make Address not roundtrip, return error when parsing signet Address. Each is absolutely terrible.