rebase due to landing D4679

In D4679#112923, @deadalnix wrote:

Could you do the same for SCRIPT_ENABLE_SCHNORR_MULTISIG and SCRIPT_VERIFY_CHECKDATASIG_SIGOPS, please?

Fully remove a stale comment (this was not removed in Core ever, but
really should have been...)

The first time I saw this PR I marked it as 'skip' in Jason's repo, but having come across it again recently, I think it has value even though we don't just do soft forks. The new discourage_upgradeable_nops is cleaner, and the new test is potentially valuable for catching some script logic bugs.

Oh actually you know what? FORKID flag is the same -- if you are doing IBD and are pre BCH activation and someone unsolicitedly sends you a BCH sighash transaction, it will create that "BUG! PLEASE REPORT THIS" log.

By the way it is possible to have a flag in extraFlags that counts as mandatory, but we need to stop using the over-simple MANDATORY_SCRIPT_VERIFY_FLAGS system and have thing specified more clearly.

Dunno why, but this was marked 'skip' in jason's tag repo.

In D4673#113120, @Mengerian wrote:

@markblundeberg you say " It is currently the only additive flag so it's the only one with weird behaviour like this."

This makes me wonder if it would be worth inverting the logic, and having a restrictive flag instead. This could also maybe be combined with refactoring NULLDUMMY flag behavior.

update comment

green with some comments

I see, my strategy was to avoid the kind of refactoring done in D4641 and D4643 and this one, and have a lighter touch to avoid too much code ownership. But, this works too.

add some regression tests

This seems to just remove functionality and Core only did this because they have problems telling apart segwit txns from partially constructed non-segwit transactions (see discussion on PR). This isn't a crutch that we need to be copying.

rebase

rm assert & update comment

rebase to master and tweak a few comments

a number of updates

• added a couple of behaviours that are nonstandard to fund but the wallet will pick up, and will spend successfully.
• rename test, the general idea is to make sure wallet is not DoSable.

update ctor

Ah OK it is documented here at least https://pubs.opengroup.org/onlinepubs/009695399/basedefs/sys/select.h.html

In D4453#107542, @jasonbcox wrote:

I think it would be better to replace this with the next upgrade, even if it does nothing for now. Otherwise, we have to re-implement this later.

remove spurious newline and request teamcity rerun

TODO:

• add SCHNORR_MULTISIG to mandatory flags
• remove NULLDUMMY completely from codebase

I find it amusing that there didn't need to be a main() before :D

bump (rebased)

rebase & revive

Going to pull this off review queue for now, since the intent is to delay landing this until some time later.

oh goody, I was looking forward to this one :)

Would have been good to have a release note for this, I just chatted with one person who was confused their regtest setup mysteriously stopped working upon upgrade.

ok, still feels weird to me to have chainwork not match the assumevalid block though

rebase for landed parent

add nonnegative check per Fabien's suggestion

update per comments

resubmit from arc (with lint changes)

This is a variant of D3818 that tests all the forms, and does some extra testing on each one.