31.0 RC Testing Guide Feedback #34827

svanstaa commented at 8:18 PM on March 15, 2026: none

This issue is to discuss the 31.0 Release Candidate Testing Guide. If you have any feedback on the document, please leave a comment here.

Note: This is for feedback on the document, not on Bitcoin Core or on the 31.0 changes. Create a new issue or give feedback in #34840.

Thank you for taking a look at the guide and leaving your feedback.

v31.0rc1 binaries are available here: https://bitcoincore.org/bin/bitcoin-core-31.0/test.rc1/.

carloantinarella commented at 5:55 PM on March 16, 2026: contributor

I think in section 6 expected results should be:

Bitcoin Core RPC client version v31.0.0rc1
Bitcoin Core daemon version v31.0.0rc1 bitcoind
Bitcoin Core RPC client version v30.2.0
Bitcoin Core daemon version v30.2.0 bitcoind

svanstaa commented at 6:55 PM on March 16, 2026: none

This is mainly for checking that the references work, point to the correct dir, and the correct version. As a good tester however, you should have an eagle eye for even the smallest deviations from expected behavior - since they might be indicative of hidden bigger problems. So I'll check it the next time working on that document, and fix it :)

Update: fixed to actual output.

carloantinarella commented at 7:55 PM on March 16, 2026: contributor

I ran the complete routine and the version output is the only deviation I found (not relevant actually, but someone completely new to the project might feel puzzled). All the rest is matching perfectly, that means you did a great job.

fanquake added this to the milestone 31.0 on Mar 17, 2026

fanquake pinned this on Mar 17, 2026

ViniciusCestarii commented at 2:17 PM on March 26, 2026: contributor

I just followed the guide and tested everything. It's a great guide!

I noticed a few things:

In the 1.3 section the script is failing with CURR_TXID: unbound variable. It's caused because the variable set before the loop is CURRENT_TXID, but the loop's createrawtransaction call references $CURR_TXID (different name).
In the 1.2 section getmempoolcluster output shows the two chunks with the same chunkfee, but the child was sent with fee_rate=2 vs the parent's fee_rate=20. Running the comamnd shows that the child's chunkfee is 0.00000282, not 0.00002820.
On Verify Setup the docs still states the wrong output. It seems that the fix wasn't applied yet.

svanstaa commented at 6:55 AM on March 27, 2026: none

Thanks , fixed in dc658c9d

itamarNov commented at 10:29 AM on April 2, 2026: none

Hi, I am relatively new to Bitcoin, software testing, and even Linux, but I used Fedora Linux 43 (Workstation Edition) to follow your guide. I’m happy to report that I successfully ran all the tests and confirmed the results! 👍 Your guide was very helpful.

Based on my experience, I have a few remarks and suggestions:

Test Automation: I initially expected the testing process to involve a single main script rather than a series of manual, step-by-step terminal commands.

Test Volume: For an application as critical as Bitcoin, I was surprised that the Release Candidate (RC) included only a few specific tests rather than a more massive suite.

Environment Persistence: Setting up the environment manually each time I opened a new terminal was tedious. It would be more efficient to provide a setup script to handle these configurations automatically.

Cluster Mempool (Cluster Limits Enforcement): I struggled with section 1.3. When I tried to copy and paste the transaction generation script directly into the terminal, the session crashed/closed. I eventually resolved this by creating a .sh file. Additionally, the script didn't recognize the bcli31 alias, so I had to explicitly define it within my script: alias bcli31="$BINARY_PATH_31/bitcoin-cli -datadir=$DATA_DIR_31"

Version Updates: While I was testing, rc1 was replaced by rc2. I wasn't sure of the standard protocol for this, so I decided to start the entire process over with rc2.

janb84 commented at 1:19 PM on April 2, 2026: contributor

Hi, I am relatively new to Bitcoin, software testing, and even Linux, but I used Fedora Linux 43 (Workstation Edition) to follow your guide. I’m happy to report that I successfully ran all the tests and confirmed the results! 👍 Your guide was very helpful.

Hi welcome ! thanks for participating.

Based on my experience, I have a few remarks and suggestions:

Test Automation: I initially expected the testing process to involve a single main script rather than a series of manual, step-by-step terminal commands.

This is more like a User Acceptance Test (UAT) and is extra, next to the extensive test suite bitcoin has. The test suite runs on every PR see our CI, example of a merge run

Next to the CI, we run Fuzzing tests (dashboard) and Mutation tests (dashboard).

Developers (and you if you have used the source code option) can run functional and Unit tests

Cluster Mempool (Cluster Limits Enforcement): I struggled with section 1.3. When I tried to copy and paste the transaction generation script directly into the terminal, the session crashed/closed. I eventually resolved this by creating a .sh file. Additionally, the script didn't recognize the bcli31 alias, so I had to explicitly define it within my script: alias bcli31="$BINARY_PATH_31/bitcoin-cli -datadir=$DATA_DIR_31"

Yes this test is a bit flaky, sorry for that. We opted to keep it in because it shows the new limit nicely (if it works). If the copy and paste would have worked, the aliases would have been recognized. Glad you figured it out !

Version Updates: While I was testing, rc1 was replaced by rc2. I wasn't sure of the standard protocol for this, so I decided to start the entire process over with rc2.

Thanks !

If you want to participate more, we have a PR review club on the 8th covering this release.

fanquake commented at 7:38 AM on April 13, 2026: member

Closing this for feedback now, as we are about at the end of the v31.0 testing period.

fanquake closed this on Apr 13, 2026

fanquake unpinned this on Apr 13, 2026