this seems to overlap with create_fund_and_spend_tx . This test is very hard to follow because it's clearly the other test with some pieces retrofited.

This also fails at the Liskov substitution principle.

http://craftedsw.blogspot.com/2010/06/liskov-substitution-principle-lsp.html

Use format(): "-replayprotectionactivationtime={}".format(ACTIVATION_TIME)

Move the comment above

Not needed, already achieved by make_conform_to_ctor

self.nodes[0] is not a function

The test seems to be doing the right checks, but starting this point, I started having a hard time reviewing it.
I found it really difficult to remember which transaction goes in which block, and jumped back and forth in the file all the time to figure it out. I would like to make some suggestions to make this easier to read:

1. next_block() and update_block() both return the block. You can get the forkblockid and postforkblockid from there, rather than retrieving them by rewinding the chain through RPCs. This has the advantage that the variables are defined at the same place where the transactions goes into the blocks, keeping related things together.
2. You can use some ascii-art in the comments to state which block contains which txns, or maybe just the mempool content with something like before [] =>after []. It is somewhat done in some comments but is not generalized.

I don' think GetBlockScriptFlags accepts null as an argument.

You can keep the message here as this is effectively what we are doing at this point in time. The fact that things will be reprocessed later is independent of what we are doing here.

dito, there is no reprocessing going on here.

This is where the reprocessing is happening. If you think is is required, this is where the message should go.

Thinking about it, I think it's fine to just say the mempool is cleared. That doesn't prevent to reinsert things at a later stage if appropriate :)

If you mock time, it's better to go on based on this rather than use time(), which make tests non reproducible.

dito

Use a loop like at line 244.

bool isEmpty() const {

remove. you literally checked the mempool the line before this one

remove; also duplicated above