Check out the debug logs.

various:

• rebase on now-landed D2479 (+ address conflicts in overlap)
• now fully mock-timed
• added actual mininode testing of non-banning (borrowed from D2479)
• added a reorg test at end, for good measure
• nits

@dagurval points out, maybe we want to go straight for https://github.com/bitcoin/bitcoin/pull/14670 ... sounds like this exits uncleanly in some cases?

some tiny nits that @Fabien pointed out on D2483 / D2527

• fully mock test time
• one more sendrawtransaction rejection test
• misc changes

rebase after D2527 landing

commandeering this one for the home stretch

rebase after D2527 landing

last nit for % formatting

address jason comments

address requests for change:

• misc nits
• logging messages altered
• test now fully mock-timed
• expanded comments about how transactions are tracked in invalidateblock and reorg events

Gonna take control of this as florian will be away for the next few days, and this needs landing soon. Thanks for the hard work thus far @florian . :-D

@Fabien ping

fix nits that Fabien pointed out

rebased; shorten comment

(make check and all functional tests passed on my local machine as well)

rebased ; release notes now applied to 0.19.0

I'm gonna grab this just to make sure it can get landed soon. Sorry for barging in!

@Fabien are these still on the table for 0.19?

Looks good on the functional test, you even did "Send post-fork-only txn into the mempool again" 👍

address comments; fix up variable names

Quick comment for now, I'd like to take a closer look later.

Looks like this is ready to go -- just return it for review from the comment Action menu.

ok; addressed my comments

(oops -- my renaming in an earlier rev had renamed too many things)

rebased onto D2527 using its mempool mechanics

ok, i liked the other one better anyway

@florian is working on a new system for the mempool reprocessing. This diff will benefit from that.

Yikes -- seems like no harm in reverting if it was good before.

OK, just one more change needed really but indeed I suppose it looks cleaner.

Change in the right direction with good tests.

In D2483#57852, @Mengerian wrote:

In D2483#57845, @markblundeberg wrote:

switch to generic flag-based activation that applies for all upgrades.

Interesting, if I understand correctly this would also work for the Replay Protection reorg code.
So does that mean the "IsReplayProtectionEnabled" portion could also be removed?
(I guess best done in a separate Diff)

A couple of comments not addressed as I disagree with them (see replies above).

switch to generic flag-based activation that applies for all upgrades.

rebased for childrens' sake

make sure peers don't ban each other when tx would be valid before/after upgrade.

clean up last revision

(still need to add that p2p banning logic)

don't clear mempool but instead reprocess it (both upgrade/downgrade)

Alternative implementation in its own file, here : D2501

rebased (fixed merge conflict with other added script_tests)

rebased ; misc nits

@dagurval I'm working on a very similar thing in D2483, perhaps our efforts would benefit from cross-pollination here.

Seems fine overall, perhaps premature though, do you have an immediate need for these refactors?

naming changes; add comments

(still need to add that p2p banning logic)

advance/rewind logic

oops, fixed the wrong thing

In D2487#57038, @Mengerian wrote:

Typo in title:
implemenation -> implementation