rebase; increase test coverage and add comments.

updated to use CheckMinimalPush

arc update to make it follow standard procedure

oops! that was me :)

just noticed this bug on testnet in my -debug=net logfile

Just to follow up, this did indeed break my testnet node's ability to sync to the fork testnet. I had to put minimumchainwork=0 in the bitcoin.conf so it would stop disconnecting imaginary_username's node with reason "Disconnecting outbound peer 51 -- headers chain has insufficient work"

@Fabien Note for reference: https://github.com/bitcoin/bitcoin/pull/13241 -- is that backport skippable now?

rebase for conflict (standard.cpp in D3807)

rebase for landed parent

move comments up a line

braces

Note -- it only supports up to 20 pubkeys like regular multisig, rather this static assert is only here since the implementation's usage of a 32-bit number (and >> operator) is implicitly assuming < 32 keys.

obsolete

Hmm, I don't get why there is the 'should' / 'must' distinction here, but yeah 👍 I wondered why it had been written that way.

rebase onto the updated parent

In D3474#88444, @teamcity wrote:

Build Bitcoin-ABC / Diffs / Diff Testing failed.

update feature_nulldummy test to only run pre-activation

Hmm, I tried also running ./test/functional/test_runner.py --with-gravitonactivation which actually causes a failure in feature_nulldummy.py, which is not surprising because the reject reason will have changed.

@deadalnix ok to commandeer now

add more test cases

In D3588#86855, @jasonbcox wrote:

This is clearly missing at least one (if not more) backports. See ReadRawBlockFromDisk in the original PR.

rebased & change to the new bitfield decoding scheme

In D3589#86858, @jasonbcox wrote:

system.h/cpp needs to be backported first

for include order (and rebuild)

Note that rejected and reconnect_p2p don't actually get used in this test; could probably be taken out. Initially I was thinking they might come in useful later, now I can't imagine that anything like that would be added.

reworked into parents

reworked parents

update comments & remove unneeded empty line

guessing at what types are expected

minor conflict -- one braces fix now done in ancestor

minor conflict from parent

rebased parent & new abc ver

address nits

update per jason's comments

[Errno 28] No space left on device

Confirm mainnet 000000000000000004425aeeb553758650da20031643fbfff0f53ac5dd1b39c3 and its chainwork; has currently 45 confirmations.

retest rebase

developer notes were done in D2941

see D3638

@deadalnix I've thought over more and I'm not convinced at all about converting to size_t. We can continue to debate but the longer this goes on, the more chance it means that dependent diffs will be delayed, resulting in Schnorr multisig being postponed until May or a later upgrade. That's fine, there's no rush to get these things done, but it seems unfortunate that feature work would be blocked on a nit like this.

include comment (except part about hashtype byte) into sigencoding.cpp

In D3581#85393, @teamcity wrote:

Build Bitcoin-ABC / Master / Gitian Builds / Bitcoin-ABC IBD Diff Testing (w/o checkpoints) passed.