Hello ๐ folks, Iโm Kevkevin. Iโm an open-source developer and reporter for Insider Edition. Last week I reviewed several pull requests from the Bitcoin Core repo. CreateNewBlock either returns a block template or throws. The null checks on that unique_ptr never ran. The converted pimpls that are always constructed are NotNull now. MuSig2SecNonce can still be moved, and nothing reads it after the move. In the mining IPC test, the createNewBlock cooldown is timed with time.monotonic(), and waitTipChanged waits at least a minute for the next block.
Thursdayโs #bitcoin-core-dev meeting stayed on the working groups. Fuzzing and QA had nothing new. Kernel has a few open test pull requests under #36692. QML is down to desktop notifications before staging. Address relay has ideas for the address-refresh spam that expired over the weekend.
Merged PRโs
Every week, several changes are officially added to Bitcoin Core. This week, 17 changes were merged. Here are some I found interesting this week.
Call sites that held a smart pointer which was never null used a nullable type, which left a null check the compiler could not delete and an error branch that never ran. CreateNewBlock has returned a block template or thrown since it landed, so the null return was dead from day one. Those checks and their error branches are gone. The return is NotNull.
The converted pimpls are constructed with the object and are not reset. The one that can be moved is MuSig2SecNonce. Nothing reads the nonce after the move. The wait path still tests for null before it wraps the template. A question about the pimplโs m_internals in validationinterface.h stayed a follow-up.
test: fix waitTipChanged timeout on slow hosts and use monotonic clock in interface_ipc_mining by Vinicius Cestari
The mined-block mining.waitTipChanged in test/functional/interface_ipc_mining.py used default_ipc_timeout. On a slow host that wait expires before the block is mined. long_timeout was already max(self.default_ipc_timeout, 60000.0) and was only used for interrupt(). The mined-block wait now uses it, so the timeout is at least 60 seconds. This failure is not a wall-clock bug: faketime reproduces it by making the test slow.
The other change replaces time.time() with time.monotonic() for the createNewBlock cooldown, whose lower bound is 0.9s, and for the duration measured after the tip wakes the test. hebastoโs NetBSD log failed with AssertionError: -14.01 < 0.9 because the wall clock stepped backward. A monotonic clock does not. waitNext on the fee-increase path has the same race, and this change does not touch it. That case landed in #36862, which also merged this week.
There are always changes being updated and reviewed in real-time. Here are some notable PRโs that are still up and looking for reviews.
It is part of #36692. sedited pointed the kernel update at the open test pull requests that name 36692. This constructor is the production API those tests are moving toward. vin and vout are generic input ranges. The constructor copies or moves the elements, not the container. Passing the vector by value and then moving the container would leave the caller a reference to an element. A later write through that reference changes the transaction after construction. The body marks that other shape with backdoor = ...; // <- modify tx, potentially breaking its invariant.
It is a conscious decision that vin/vout are passed as generic input ranges and that only their elements are copied/moved.
IRC meeting notes
Every week on Thursday, there is an IRC meeting. Here are some short notes from that meeting.
๐ณ๐ท๐ฎ๐ต๐ฟ: #bitcoin-core-dev Meeting: _aj_ abubakarsadiq achow101 andrewtoth b10c brunoerg cfields danielabrozzoni darosior dzxzg eugenesiegel fanquake fjahr furszy hebasto hodlinator instagibbs janb84 jarolrod johnny9dev jonatack josie jurraca kanzure kevkevin laanwj lightlike l0rinc maflcko marcofleon maxedw Murch pinheadmz provoostenator pseudoramdom ryanofsky sdaftuar sedited sipa sliv3r__ sr_gi stickies-v stringintech theStack vasild willcl-ark
๐ณ๐ท๐ฎ๐ต๐ฟ: There are no pre-proposed meeting topics this week. Any last minute ones to add?
๐ณ๐ท๐ฎ๐ต๐ฟ: Letโs get started with the WGs
--- Topic 1 ---
๐ณ๐ท๐ฎ๐ต๐ฟ: #topic Fuzzing WG Update (eugenesiegel, marcofleon)
๐ฒ๐๐ด๐ฒ๐ป๐ฒ๐๐ถ๐ฒ๐ด๐ฒ๐น: no update
--- Topic 2 ---
๐ณ๐ท๐ฎ๐ต๐ฟ: #topic Kernel WG Update (sedited)
๐๐ฒ๐ฑ๐ถ๐๐ฒ๐ฑ: there are a few pull requests open aiming at improving tests https://github.com/bitcoin/bitcoin/pulls?q=is%3Apr+36692+in%3Abody+is%3Aopen
๐๐ฒ๐ฑ๐ถ๐๐ฒ๐ฑ: thatโs it for now.
--- Topic 3 ---
๐ณ๐ท๐ฎ๐ต๐ฟ: #topic QML GUI WG Update (johnny9dev)
๐ท๐ผ๐ต๐ป๐ป๐๐ต๐ฑ๐ฒ๐: Spent most of last week working with my agents on resolving the 300 issues reported
๐ท๐ผ๐ต๐ป๐ป๐๐ต๐ฑ๐ฒ๐: Resolved+test: 248 | Pre-PR fixed: 7 | Unconfirmed: 52 | Deferred: 15 | Open: 7 (Total: 329)
๐ท๐ผ๐ต๐ป๐ป๐๐ต๐ฑ๐ฒ๐: those are the numbers
๐ท๐ผ๐ต๐ป๐ป๐๐ต๐ฑ๐ฒ๐: Iโm currently working on deslop and cleanup. Its getting close to looking really good
๐ท๐ผ๐ต๐ป๐ป๐๐ต๐ฑ๐ฒ๐: The last remaining feature to complete is desktop notifications. So with issues resolved, desktop notifications, and current PRs merged in weโll be ready to do staging
๐ท๐ผ๐ต๐ป๐ป๐๐ต๐ฑ๐ฒ๐: that is all
--- Topic 4 ---
๐ณ๐ท๐ฎ๐ต๐ฟ: #topic QA WG Update (brunoerg)
๐ฏ๐ฟ๐๐ป๐ผ๐ฒ๐ฟ๐ด: no update this week
๐ณ๐ท๐ฎ๐ต๐ฟ: I didnโt see any of the other wg leads, anyone still lurking or did I miss someone?
๐ณ๐ท๐ฎ๐ต๐ฟ: Otherwise is there anything else to discuss?
๐ฏ๐ญ๐ฌ๐ฐ: stratospher might not be here. We looked into some of the address-refresh spam expiring over the last weekend. Have some ideas for fixes. More soon.
๐ฏ๐ญ๐ฌ๐ฐ: (for address relay)
๐ณ๐ท๐ฎ๐ต๐ฟ: b10c: Thanks!
Releases
Two maintenance releases have binaries, tags, and release notes dated 19 Nov 2026. The releases index on bitcoincore.org and the GitHub Releases list have not been updated. The newest GitHub Release object is still v32.1 from 28 Oct 2026. There is no 32.2 directory. #36727, the 32.x backport, did merge this week.
Bitcoin Core 30.5. Prep is #36893. Release notes credit fanquake, Greg Sanders, and willcl-ark. The net change is #35878 from Greg Sanders. Conflicting advertisements for a transaction are not both processed. They cannot both be accepted into the mempool, so the extra work is skipped. The pull request says there should be no change in observable behavior. CI notes on this branch are #33693, #35348, #35441, and #35571.
Bitcoin Core 31.3. Tag v31.3 is 107ecee166a4ddf7acb8d9cc9b8627888becea91. Prep is #36892. The same #35878 net change is in the notes. Credits are Greg Sanders and willcl-ark. CI notes are #35441 and #35571.
Thank you for reading. Be sure to tune in again next week for your updates on Bitcoin Core!
If there are any comments, suggestions, or errors, do not hesitate to reach out or comment.


