Again, the issue is caused by --descriptors causing a different fee calculation, so the balance on node0 is not exactly the same.

partial fix for wallet_keypool, add some comments when our codebase diverges for a known reason.

3 more tests to fix: wallet_keypool (with and without --descriptors), wallet_basic (with --descriptors) and rpc_psbt (with and without --descriptors)

Ok I actually read the comments. If there is a rationale for the previous approach then let's stick with that, I don't know if there is a plan to remove that function in the future.

rebase onto master.

Try to get this fixed. This is now blocking a couple of backports.

use [[maybe_unused]] when defining the function instead of planting dummy references to the function

update the description to add a note about how the tx fees are computed, and add a comment next to -maxapsfee=2.93 to explain the threshold.

In D10091#227354, @Fabien wrote:

Is that a backport ?

In D10090#227352, @Fabien wrote:

What about [[maybe_unused]] ? That would be more elegant.

I will try to figure out exactly why the second set of values, the computed fees for the transactions created in the test, are different.

TIL

Please take care to use the appropriate test plan. This one was not good (I edited), but luckily the CI did the job.

@bot build-tsan

Is that a backport ?

What about [[maybe_unused]] ? That would be more elegant.

Unrelated failure

Improve RPC help