< bitcoin-git>
bitcoin/master ad57fb7 Wladimir J. van der Laan: wallet: Add BerkeleyDB version sanity check at init time
< bitcoin-git>
bitcoin/master 80486e7 Samuel Dobson: Merge #20952: wallet: Add BerkeleyDB version sanity check at init time
< bitcoin-git>
[bitcoin] meshcollider merged pull request #20952: wallet: Add BerkeleyDB version sanity check at init time (master...2021-01-berkeleydb-sanity-check) https://github.com/bitcoin/bitcoin/pull/20952
< bitcoin-git>
[bitcoin] fujicoin opened pull request #20970: Fix a bug that personal signet can not be created. (master...fix-personal-signet) https://github.com/bitcoin/bitcoin/pull/20970
< bitcoin-git>
[bitcoin] MarcoFalke closed pull request #20970: Fix a bug that personal signet can not be created. (master...fix-personal-signet) https://github.com/bitcoin/bitcoin/pull/20970
< hebasto>
prayank: fyi, also ##bitcoin-core-gui exists
< jonatack>
vasild: iirc it's one of the issues we discussed while testing together, and another tester from twitter managed to follow the testing setup steps i gave here https://twitter.com/jonatack/status/1351136998854176777?s=20 and reported it too
< vasild>
is it the one where the rpc "addnode ... remove" does not disconnect the peer immediately?
< jonatack>
it surprises people that addnode onetry doesn't count for addnode remove
< vasild>
this behavior is not specific to I2P, same applies to IPv4 addresses on master
< jonatack>
"Node could not be removed. It has not been added previously."
< jonatack>
right, not specific to i2p
< jonatack>
maybe the error message could be clearer
< vasild>
two "surprises": 1. "addnode foo onetry" followed by "addnode foo remove" produces an error and 2. "addnode foo add" connects to the node, but then "addnode foo remove" does not disconnect from foo
< vasild>
both made sense to me once I understood how it works :) maybe some docs/error messages improvement is warranted indeed
< jonatack>
yes. disconnectnode works, addnode remove was either slow or no effect, i don't recall
< vasild>
"addnode foo remove" removes foo from the list of "keep connections to these nodes". After being removed from that list foo becomes a normally connected peer, subject to eivction and we will likely not connect to foo after restart. But definitely we will not disconnect from foo immediately due to "addnode foo remove"
< jonatack>
oh right, i hadn't looked at the addnode help in a while, and looking at it now, it seems ok as-is but a bit more explanation might not hurt
< vasild>
jonatack: https://bpa.st/LTDBA I connected two times to your bitcoin node via i2p on different ports
< vasild>
probably now you see two incoming connections from me
< vasild>
that is something that needs to be addressed
< jonatack>
vasild: yes, saw double conns a couple of days ago, same peer, two peer ids, lasting for a short time https://imgur.com/a/IaFfjCx
< vasild>
your port is not listed in that image, may be the same thing or may be different
< jonatack>
right now i don't see your inbound i2p peers because i have today's review club PR #20546 built, will reload #20764 after the meeting
< vasild>
hmm, so it must be different thing then - I managed to connect to you twice by using a different port, but medea connected to you twice because...?
< wumpus>
jonatack: agree wrt 20951 (for context, the person who opened that PR first mailed me personally about the suggestion, i told them to open a PR instead)
< wumpus>
it... doesn't seem like a big improvement of the text, at most a very subtle difference, not sure it warrants a post-hoc release notes change
< wumpus>
but it's up to kallewoof
< benthecarman>
#proposedmeetingtopic Thoughts on a future IRC meeting discussing taproot activation?
< luke-jr>
hebasto: re GUI#186, I suspect the ideal solution would be to instead re-open the send in the Send tab editor-style
< gribble>
https://github.com/bitcoin/bitcoin/issues/186 | Remove unnecessary line which creates a dependency on wxs png library. by TheBlueMatt · Pull Request #186 · bitcoin/bitcoin · GitHub
< luke-jr>
hebasto: then coin control users can select/deselect inputs (with a warning if they try to deselect all common inputs to replaced transactions)
< luke-jr>
step 1 might be to start saving a flag in the wallet with CC options
< hebasto>
luke-jr: sounds good as a more general solution
< pinheadmz>
is BIP125 (RBF) accurate with the implementation? Specifically, is it the absolute fees that are compared between replaces and replacing tx? or the fee rate?