Since it's coming from a different repo than most ABC backports, the Diff summary should mention what repo these backports are coming from and/or link to them

Fix typo

Add link directing contributors to Bitcoin ABC Phabricator repo
Re-add Travis link, now that Travis is enabled for the repo

Add a bit more explanation of how the repository is maintained, so that people can understand what's going on.
Re-add "Use at your own risk" disclaimer as recommended by @Fabien
Add link and mention that this is the Bitcoin Cash variant of the Schnorr algorithm.

@markblundeberg It may also be good to add some Schnorr information to the "Implementation Details" portion, and maybe a link to the BCH Schnorr spec on bitcoincash.org. Please let me know if you have any suggestions.

In D4673#113173, @markblundeberg wrote:

Also, what is NULLDUMMY? No such flag exists. ;-)

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

Good catch!

Great to see this!

Looks good to me!

Would be good to also update CONTRIBUTING.md with this version.

Looks good. Based on the crawler at cash.coin.dance/nodes there seem to be enough 0.20.x nodes out there to make this change now.

Rebase on updated D3972

checked on my local node, looks good.

Looks good to me.

Chainwork is above the value on the Upgrade testnet, which will make it more complicated for people to run this release on the upgrade testnet.

Seems good, but pretty soon we should update the version number in contrib/seeds/makeseeds.py

FYI, Updating testnet params would be fine this time, as we will be re-starting the Upgrade testnet with this release.

Rebase

Looks good.

Looks fine to me.

Next version should be 0.20.0 no?

In D3830#90139, @jasonbcox wrote:

The chainwork should be as high as the highest seen PoW and does not need to match the assume valid block. However, it should always be greater than the assume valid block, which is obvious. See D3732

Seems strange that chainwork does not correspond to the block hashes used for assumevalid.

Wording now looks good to me

@Fabien @markblundeberg How about this: "Schnorr multisig checkbits implementation assumes < 32 pubkeys."

This makes the code more readable for me, so it seems like a good change.

The logic looks good to me.

I would suggest updating the description, which refers to "ints".

Oh, it is changed in master :)

Someone should change doc/release-process.md to correspond to this new set of places that the version number is updated.