< bitcoin-git> [bitcoin] BJBRbygJtzBfp4gjJG2iqL opened pull request #19834: Update bitcoin_config.h (master...master) https://github.com/bitcoin/bitcoin/pull/19834
< bitcoin-git> [bitcoin] fanquake closed pull request #19834: Update bitcoin_config.h (master...master) https://github.com/bitcoin/bitcoin/pull/19834
< fanquake> wumpus / sipa: can you block BJBRbygJtzBfp4gjJG2iqL
< sipa> fanquake: done
< gleb> sdaftuar: I was just going through the p2p logs I couldn't attend and I think we have some overlapping ideas. I see 2 dimentions in "small topology improvements": 1. talking to new peers (not necessarily rotation, feelers too) and exchanging new data (sync tips)
< gleb> It seems you're planning to work on syncing tips with feelers soon. I wanna go even further and a) introduce new temp connections: talk to the historically good peers. This is somewhat related to anchors as it addresses a similar attack vector, but perhaps not making them persistent would be a good tradeoff.
< gleb> Another idea is to make self-announcements go to these feelers/temp_conns, perhaps with some probability. I think this makes us prone to some Addr relay DoS and obfuscates addr relay even further.
< bitcoin-git> [bitcoin] MarcoFalke opened pull request #19836: rpc: Properly deserialize txs with witness before signing (master...2008-rpcDeserTxsWitness) https://github.com/bitcoin/bitcoin/pull/19836
< bitcoin-git> [bitcoin] hebasto closed pull request #19668: Do not hide compile-time thread safety warnings (master...200805-assert) https://github.com/bitcoin/bitcoin/pull/19668
< bitcoin-git> [bitcoin] hebasto reopened pull request #19668: Do not hide compile-time thread safety warnings (master...200805-assert) https://github.com/bitcoin/bitcoin/pull/19668
< hebasto> it seems appveyor is broken
< hebasto> sipsorcery: ^
< sipsorcery> ok will take a look, the vs2019 image got updated yesterday so that's prob the cause.
< hebasto> sipsorcery: thanks!
< jonatack> yup, appveyor is down for me too, thank you hebasto for bringing it up and sipsorcery for having a look
< bitcoin-git> [bitcoin] sipsorcery opened pull request #19839: Set appveyor VM version to previous Visual Studio 2019 release (master...appveyor_fix) https://github.com/bitcoin/bitcoin/pull/19839
< gzhao408> Could someone help restart this Travis job that errored? https://travis-ci.org/github/bitcoin/bitcoin/jobs/722470895
< fjahr> done
< gzhao408> fjahr thank you!
< bitcoin-git> [bitcoin] promag opened pull request #19840: Avoid callback when -blocknotify is empty (master...2020-08-blocknotify) https://github.com/bitcoin/bitcoin/pull/19840
< gzhao408> Another request to restart Travis job 😅 https://travis-ci.org/github/bitcoin/bitcoin/jobs/722470890
< bitcoin-git> [bitcoin] sipa opened pull request #19841: Implement Keccak and SHA3_256 (master...202008_sha3) https://github.com/bitcoin/bitcoin/pull/19841
< kallewoof> wumpus: sorry for late response; I tried adding 'const' but warning persists; boost::{anonymous}::none can't be constexpr'd, it looks like, but I could be getting the syntax wrong.(error = "the value of ‘boost::{anonymous}::none’ is not usable in a constant expression")
< kallewoof> wumpus: It could be that the gcc version is too old though; the machine giving the warnings is gcc 5.4.0, and warnings do not appear for 7.5.0
< sipa> kallewoof: context?
< kallewoof> sipa: one of my machines gives ./optional.h:24:14: warning: ‘nullopt’ defined but not used [-Wunused-variable] for "static auto& nullopt = boost::none;"
< kallewoof> i realized just now that its gcc is older than the others, so the warning can probably be ignored.
< kallewoof> i dont see that error on other machines, only this one.
< kallewoof> (i should add that it gives that error for every #include, so like 40 or so times)
< sipa> seems harmless
< sipa> i assume it's some issue with an old gcc, indeed
< kallewoof> Yeah, I was mostly curious about why 40+ warnings were ignored by the devs (now I know why; outdated compiler)