test: automatically pick bitcoind or bitcoin-node

[Sjors]

Testing bitcoin#33229 with an additional test simplification.

[Sjors]

@willcl-ark any idea why MSan and TSan would fail on my fork? I'm not using the paid runners, so IIUC since bitcoin#32989 it should fall back to free Github runners. I could try on to see if that's the culprit.

#9 232.9 fatal: Remote branch llvmorg-21.1.1 not found in upstream origin
#9 232.9 Retries exhausted
#9 ERROR: process "bash -c cd /ci_container_base/ && set -o errexit && source ./ci/test/00_setup_env.sh && ./ci/test/01_base_install.sh" did not complete successfully: exit code: 128

https://github.com/Sjors/bitcoin/actions/runs/17578075963/job/49927909949?pr=104

I opened #105 to further debug CI on a more recent master commit.

[ryanofsky]

This is nice idea. A change like cf08bc9 would be good so tests just use the bitcoin wrapperand and test IPC by default without CI configurations needing to specify BITCOIN_CMD. I think some changes might be need to be made to respect BITCOIND variable though, so if a BITCOIND path is set it will be used instead of the bitcoin wrapper, to not break developer workflows and CI jobs that may be using BITCOIND. It may also make sense to keep BITCOIN_CMD option around too.

[willcl-ark]

@willcl-ark any idea why MSan and TSan would fail on my fork? I'm not using the paid runners, so IIUC since bitcoin#32989 it should fall back to free Github runners. I could try on to see if that's the culprit.

#9 232.9 fatal: Remote branch llvmorg-21.1.1 not found in upstream origin
#9 232.9 Retries exhausted
#9 ERROR: process "bash -c cd /ci_container_base/ && set -o errexit && source ./ci/test/00_setup_env.sh && ./ci/test/01_base_install.sh" did not complete successfully: exit code: 128

Sjors/bitcoin/actions/runs/17578075963/job/49927909949?pr=104

I opened #105 to further debug CI on a more recent master commit.

This appears to be related to bitcoin#33345 (now fixed by a tag appearing upstream).