SPV, RPC, Rescan fixes

[rithvikvibhu]

SPV, RPC and rescan fixes. Redid commits to make review easier.

node: set spv boolean correctly

• closes Needs app restart on toggling SPV mode #467
• Now there's no need to restart Bob multiple times when toggling SPV.

ui: hide generatenewblocks if spv

• closes "generate to address" buttons in regtest won't work with SPV node #466
• SPV can't generate blocks, hides the +1, +10, +50 buttons

node: remove getAuctionInfo

• Get Auction Info is a wallet call, but was part of node service, removed it

node: common api url constant

• A single constant for URL for handshake api (remote node used in SPV mode) - easier to replace during dev

exchange: disable in spv mode

• Related to Shakedex will not work in SPV mode #480
• Disables Exchange page in SPV with a message. Keeping the issue open until shakedex is updated to use Bob's node service including remote api calls

rpc: allow http for private IP addresses

• From Discord, a user wanted to use Bob over RPC to a node on a local network.
• Instead of allowlisting just localhost, it now checks and allows HTTP for private IP ranges too (hostnames still require HTTPS, even those that resolve to private IP addresses).
• Uses binet (added to package.json, but hsd itself requires it so doesn't add anything)

rpc: splash screen switch to internal node on error

• closes Add the rig options on the loading screen (Working on it) #515
• If Bob is set to use a remote RPC node, and the node is not reachable, Bob errors on splash screen
• If the node cannot be started (say the IP address is released, or domain is not under user's control), then the only way to get Bob working again was to clear all Bob settings by deleting the db directory.
• Now, if it's an RPC connection failure, the splash screen has a button to switch to internal node

node: throttle refreshNodeInfo

• refreshNodeInfo is a slightly expensive call (getInfo and getFees) and was being called for every new block, which is very fast in SPV mode
• It is now throttled.

wallet: replace rescan hooks with block connect events

• closes Status stuck on rescan in SPV mode #421, closes Auction rescan broken when using custom RPC, (0.6.2) #310
• depends on wallet: emit 'block connect' events handshake-org/hsd#707, net: fixes for SPV handshake-org/hsd#752
• The biggest change of this PR, replaces the rescan unhook and unhooking mitm with wdb block connect events.
• Simplifies rescan logic that works well with Full, RPC and SPV (thanks @pinheadmz for SPV, RPC, Rescan fixes #469 (comment))
• SPV (active) auction rescans requires net: fixes for SPV handshake-org/hsd#752. Why it affects us (read this PR desc for (1) and (2):

• Before requesting blocks, a node asks for peer's mempool (but only in regtest!)
• We know that sendGetBlocks doesn't really request anything, so nothing happens after that (2)...
• BUT, because of (1), the node expects a mempool inv response - which it won't get because mempool is empty - and after 30 secs, kills the connection.
• With only set to a single peer, it connects to the same peer again and now makes getBlocks for "the first time" and gets the new blocks.
  To a Bob user, it'll look like a rescan took >35 seconds.

Now, rescans work everywhere and show proper status (next commit below):

	Settings -> Rescan	Auction Rescan
Full node	✅	✅
SPV node	✅	✅
RPC - Full node	✅	✅

wallet: correct resync progress in spv

• resyncing progress in sidebar and topbar now use resync height added in previous commit

pkg: update to hsd v4.0.1

• Updates hsd to v4.0.1 which has net: fixes for SPV handshake-org/hsd#752 merged

---

Overall, SPV seems more stable now, and syncs from scratch in 25 minutes.

TODO:

• Fix tx history not refreshing on rescan complete
• Resolve all "TODO" comments
• Import names in spv
• Update hsd version after wallet: emit 'block connect' events handshake-org/hsd#707 is merged
• Remove console.logs

[pinheadmz]

👍 looks good so far.

[pinheadmz]

@rithvikvibhu re: 7af427a

I think we can make this a lot simpler by adding a feature to hsd: handshake-org/hsd#621

If wallet and node both emit "new chain height" events, we will always know their sync status and we can drop the unhook hack and forget about "wallet rescan" events entirely. This should work for SPV fine as well.

bob-wallet/app/background/wallet/service.js

Line 203 in de90b37

this.node.wdb.client.unhook('block rescan');

[rithvikvibhu]

@pinheadmz using hsd fork (handshake-org/hsd#707) and yeah it made things simpler!

SPV importname is still a problem though.

[rithvikvibhu]

Rebased, updated to hsd v4.0.1 and it's SO satisfying to watch the auction rescan in SPV!
The blue progress bar of the bidding period goes back to 0 and increments smoothly.