< ariard>
oh that's nice, it's interesting to see how much reads are expensive even if it's something you always expect
< kallewoof>
ariard: no the mempool is empty afterwards, it's just a big spam fest listing every coinbase ever mined by that node (i think)
< kallewoof>
This was not the case. Not sure what was different. --> <kallewoof> MarcoFalke: Hm, not sure if 16452 is related or not from looking at the changes. Anyway Ithink you can trigger this by simply starting two regtest nodes and reorging one to the other. They should start spamming those messages I think
< ariard>
kallewoof: my PR removed coinbase check in SubmitMemoryPoolAndRelay because it was already checked in ATMP, what you're saying is coinbase txn being resubmited to mempool despite being confirmed?
< fanquake>
wumpus: looks like there is also at least 5204, 4908, 3095, 2936
< wumpus>
fanquake: ooh fun
< wumpus>
fanquake: we should deprecate problematic things more often :)
< fanquake>
wumpus: I agree
< fanquake>
wumpus: I actually rebased my disable BIP70 by default PR earlier today as well.. #15584
< sipa>
emilengler: you can find these things yourself by grepping the source code for strings you see appear
< fanquake>
wumpus: yea, realised after another look, have left that open
< wumpus>
emilengler: yes, or LogPrint(BCLog::ADDRMAN, for specific categories that can be specified with -debug=...
< emilengler>
sipa: Ok, will do this in future :) I might be using this channel to often ;)
< sipa>
emilengler: it's not that you're not welcome asking questions, but you can learn a lot more by investigating things yourself
< wumpus>
yes, printing to the debug log seemed too straightforward a question, I was worried for a bit that he wanted to print, to a printer :)
< sipa>
ha
< emilengler>
I think priting is better term then putting
< sipa>
"logging"
< sipa>
transitive verb
< sipa>
1a : to cut (trees) for lumber
< sipa>
b : to clear (land) of trees in lumbering —often used with off
< wumpus>
hah i remember that the icon for gnome-logger (or a tool to browse system logs, at least) was a literal log of wood at some point, this confused people for whom English is not their native language a lot
< wumpus>
fanquake: I don't see the comment you added in #16533, maybe you pushed the wrong branch?
< wumpus>
oh, you added it only to the commit description and not in the change itself, ok, that wasn't clear to me
< * wumpus>
tends to interpret 'comment' as code comment 😄
< kallewoof>
ariard: see my pastebin link etc. I've never seen that kind of stuff before. The coinbase is obviously not getting into the mempool because it is not a valid mempool tx, but the logger is saying it attempts to do so for what it looksl ike a lot of or all of the coinbases related to the node's wallet.
< fanquake>
wumpus: right, that might actually be what dongcarl meant. I can update with a comment in the .mk file if you prefer.
< emilengler>
I think I'm using LogPrintf() wrong
< emilengler>
It logs stuff that does not belong there
< emilengler>
It logs stuff like "Env: opencon thread exit" eventhough it should only log a %s with a linebreak
< emilengler>
Ok nevermind looks like the problems are at my pointer
< fanquake>
wumpus: added both now anyways.
< bitcoin-git>
[bitcoin] fanquake opened pull request #16554: test: only include and use OpenSSL where it's actually needed (BIP70) (master...test_openssl_include) https://github.com/bitcoin/bitcoin/pull/16554
< wumpus>
fanquake: yes, i think adding both is best for these kinds of things: a source code comment is more apparent than a commit comment on casual browsing, but it's good to have the rationale in git history too
< bitcoin-git>
[bitcoin] Sjors opened pull request #16555: [doc] mention whitelist is inbound, and applies to blocksonly (master...2019/08/doc-whitelistrelay) https://github.com/bitcoin/bitcoin/pull/16555
< bitcoin-git>
bitcoin/master 2f37163 fanquake: test: only include and init openSSL where it's actually used
< bitcoin-git>
bitcoin/master e5fdda6 MarcoFalke: Merge #16554: test: only include and use OpenSSL where it's actually neede...
< bitcoin-git>
[bitcoin] MarcoFalke merged pull request #16554: test: only include and use OpenSSL where it's actually needed (BIP70) (master...test_openssl_include) https://github.com/bitcoin/bitcoin/pull/16554
< ariard>
kallewoof: ah yes now the logger in SubmitMemoryPoolAndRelay is printing before confirmation checks are done in BroadcastTransaction
< ariard>
IMO best would be to remove txn with more 100 confs from mapWallet to avoid broadcast retries of old confirmed ones