fix diff summary, implement review suggestion (return AbortNode....)

I think my summary (a copy paste of the initial message in the pull request) is misleading. The PR later change it's scope to make no change in behavior (besides additional logging) and basically just force future users of these functions to handle the return value.
I will fix this

translation nit

lgtm (if the bitcoin-chainstate / kernel build does not fail)

This whole diff is a test on how Grok v3 can helpful to do semi-automated changes on a large amount of code.
It worked quite well, I just had to revert a few changes msg_avaproofs.serialize and msg_avaproofsreq.serialize, which imo were not trivial enough to justify removing the intermadiate var and all the comments to make it a one-liner.

back-off unrelated changes

mention which devices have a custom Trezor firmware

reflow release note

use slip10 1.0.1 to better satisfy the trezorlib >= 1.0.1 requirements (the AppImage build works with 1.0.0, but with more recent versions of pip the conflict becomes an error)

update the repo and commit hash to ecash/2.8.9 tag from the Bitcoin ABC repo

factor the token detection into a wallet.has_tokens(coin) method and fix one more thing that could cause accidental spending of ALP tokens (for users who enabled "Spend only fused coins", if a token is sent to the same address as a fused coin)

In D17806#405784, @bytesofman wrote:

TIL coin consolidator is a thing in electrum

how is this used in the GUI? is this just called in normal tx building?

As an additional test I'm also going to try to be in a few CashFusion rounds to confirm no tokens are burned. But I cannot guarantee the tests will be conclusive in a timely manner.

tested with python3.11 (via AppImage)