<fanquake>
vasild: Not sure if we are going to cut an rc5 just for that
<fanquake>
Or if we are, we should just kill rc4 now
timbo_xyz has joined #bitcoin-core-dev
<fanquake>
There are other private broadcast test fixes for sporadic failures that also seem like they will miss 31.0, so these followups could go into 31.1
afiore has quit [Remote host closed the connection]
afiore has joined #bitcoin-core-dev
afiore has quit [Remote host closed the connection]
vasild has quit [Remote host closed the connection]
timbo_xyz has quit [Ping timeout: 265 seconds]
afiore has joined #bitcoin-core-dev
vasild has joined #bitcoin-core-dev
bitdex has joined #bitcoin-core-dev
timbo_xyz has joined #bitcoin-core-dev
PaperSword has quit [Quit: PaperSword]
memset has quit [Remote host closed the connection]
afiore has quit [Remote host closed the connection]
afiore has joined #bitcoin-core-dev
memset has joined #bitcoin-core-dev
Randolf has joined #bitcoin-core-dev
timbo_xyz has quit [Remote host closed the connection]
afiore has quit [Remote host closed the connection]
<bitcoin-git>
[bitcoin] theuni opened pull request #35040: Remove CNode dependency for local address functions and introduce LocalAddressManger (master...addrlocal-refactor3) https://github.com/bitcoin/bitcoin/pull/35040
janb84 has quit [Ping timeout: 248 seconds]
Emc99 has joined #bitcoin-core-dev
janb84 has joined #bitcoin-core-dev
dzxzg has joined #bitcoin-core-dev
cfields has joined #bitcoin-core-dev
<abubakarsadiq>
#startmeeting
<corebot`>
abubakarsadiq: Meeting started at 2026-04-09T16:00+0000
<corebot`>
abubakarsadiq: Current chairs: abubakarsadiq
<pinheadmz>
From fjahr: "#34158 was merged. Now my only PR left on this project is #34342 and since it depended on some changes from #34905, which was also merged, I was now able to take it out of draft status and it’s ready for review. It has some small parts shared with #32061 but I have decoupled it for now to allow for independent review."
<corebot`>
pinheadmz: Error: That URL raised <Connection timed out.>
<pinheadmz>
And for me, now that 34905 was merged, next PR is #34772 which currently has 2 ACKs. Then I can rebase #32061 (removing the first 7 commits) and address the feedback there from hodlinator and vasild.
<corebot`>
https://github.com/bitcoin/bitcoin/issues/32061 | Replace libevent with our own HTTP and socket-handling implementation by pinheadmz · Pull Request #32061 · bitcoin/bitcoin · GitHub
<pinheadmz>
and thats all for us !
<cfields>
🚀
<dzxzg>
Very cool
<abubakarsadiq>
#topic QML GUI WG Update (johnny9dev)
<johnny9dev>
focused on external signer support this week. I have configuration and wallet creation and working through getting signing setup in the send flow.
<johnny9dev>
epicleafies: any status?
<epicleafies>
was out of town, planning on working on another uri issue and fixing up previous PRs
<johnny9dev>
thanks, that is all
<abubakarsadiq>
#topic Kernel WG Update (sedited)
<cfields>
I believe he's still away
<abubakarsadiq>
sedited is not here, but I pushed #35000, to add unit + fuzz test for block validity, motivation and other details are in description.
<corebot`>
https://github.com/bitcoin/bitcoin/issues/35000 | test: Add block validation unit tests + fuzz target for `TestBlockValidity` by ismaelsadeeq · Pull Request #35000 · bitcoin/bitcoin · GitHub
<hodlinator>
hi
<abubakarsadiq>
That's it from me. I don't think there is any kernel-related update.
<abubakarsadiq>
#topic Benchmarking WG Update (l0rinc, andrewtoth)
<l0rinc>
#34208 was merged, from now on we can make a few benchmarks more realistic by excluding the expensive setup times from the measurements.
<abubakarsadiq>
There are no pre-proposed meeting topics this week. Any last minute topic?
<_aj_>
are there any docs on how long users should expect IBD to take? i thought on modest hardware it was a couple of days, but it seems to be ~12 hours?
<l0rinc>
rpi4 is very slow, but rpi5 is already less than a day with latest assumevalid
<pinheadmz>
how "modest" is your hardware ?!
<_aj_>
vm on >5 year old high-end desktop?
<andrewtoth_>
there are a lot of determining factors
<l0rinc>
after #31132 even rpi5 finishes in less than 10 hours
<abubakarsadiq>
_aj_: I dont think we have ibd.md doc
<l0rinc>
and they have SHA acceleration via the cryptographic extension
<l0rinc>
16 GB mem performs best, 8 GB mem takes 1.5 days
<abubakarsadiq>
loric: "latest assumevalid" you mean bumping to chain tip?
<sliv3r__>
l0rinc pi5 using the nvme slot?
<l0rinc>
yes, latest master
purpleKarrot has quit [Quit: purpleKarrot]
purpleKarrot has joined #bitcoin-core-dev
<l0rinc>
PCIe, not usb anymore
eugenesiegel73 has joined #bitcoin-core-dev
bugs_ has joined #bitcoin-core-dev
<andrewtoth_>
having a peer connected via local network helps as well. Using public p2p network is a crapshoot
<l0rinc>
but a lot of users still set the dbcache equal to the memory - which makes it swap and IBD will take weeks. We added warnings and documentation to avoid that
<abubakarsadiq>
andrewtoth_: we can't use that as a realistic bench for other users though, we expect people to sync via public p2p
<_aj_>
maybe benchcoin or something could have some example configs and performance stats for half a dozen common example configs?
<sliv3r__>
_aj_: maybe not docs saying how long should expect but giving hints on how to speed it up
<l0rinc>
I usually sync via real nodes when checking IBD, I just run it multiple times for stability
<_aj_>
(if p2p is the bottleneck now, we should maybe just improve that more aggressively; i've no idea how much it is, and i think .au latency tends to make it worse too)
eugenesiegel has quit [Ping timeout: 245 seconds]
<abubakarsadiq>
_aj_: +1
<andrewtoth_>
abubakarsadiq: yes, but with public p2p there is no determinism
<l0rinc>
there isn't a lot that can be configured, we've modified the related default configs to accommodate the new reality
luke-jr has quit [Ping timeout: 264 seconds]
<l0rinc>
yes, download is already the bottleneck for the average internet speed. With 100 mbps it takes ~16 hours just to download all the data.
<_aj_>
if the default config is good enough, then performance stats on common hard (rpi4, rpi5, low-end desktop, high-end desktop, cheap vps, expensive vps) would be pretty good?
<l0rinc>
_aj_ if you have any ideas which areas to investigate, please let me know, I'd love to play with it
<l0rinc>
_aj_ my M4 PR laptop can do a full validation in less than 2 hours
<andrewtoth_>
#31132 helps a lot with cheap vps, since network connected storage has high latency
<_aj_>
l0rinc: the performance i saw was great, it was just my expectations that were out of whack. i think my remote peers were slowing down block download from my local peer, but only connecting to my local peer solved that, so
<l0rinc>
abubakarsadiq: after andrewtoth's parallel input fetcher the dbcache size doesn't matter as much anymore
<_aj_>
cfields: (i did see a lot of cpu from the sending peer fwiw, so i expect that would have helped a little, i thought about turning v2transport off but didn't)
<achow101>
#proposedtopic 31.0
<abubakarsadiq>
#topic 31.0 (achow101)
<l0rinc>
cfields: I will prioritize reviewing that in more depth if you have time to work on it
<achow101>
rc4 was tagged yesterday, please test. Hoping this is the last rc
<cfields>
l0rinc: ack, thanks.
<achow101>
And please review the release notes draft in the wiki
<achow101>
That's all
<abubakarsadiq>
Anything else to discuss?
<abubakarsadiq>
#endmeeting
<corebot`>
abubakarsadiq: Meeting ended at 2026-04-09T16:29+0000