< wumpus>
luke-jr: isn't a batched request considered to be a single request from the viewpoint of the queue?
< wumpus>
luke-jr: the commands in a batch are always executed in order anyway, there would be no point trying to queue them to different threads
< wumpus>
(if anything it would break the gaurantee they'll finish in order)
< wumpus>
the work queue is to distribute incoming work over worker threads nothing more
< bitcoin-git>
[bitcoin] MarcoFalke opened pull request #20146: net: Send post-verack handshake messages at most once (master...2010-netPostVerackHandshake) https://github.com/bitcoin/bitcoin/pull/20146
< jnewbery>
hmmm, #19988 is no longer showing up in my notifications stream in github. I can still open the PR and see it in the list of open bitcoin/bitcoin PRs
< jnewbery>
I don't think I silenced notifications for that PR
< aj>
jnewbery: still shows up for me
< jnewbery>
it showed up in 'Done' category. I generally don't mark notification in github as Done, but maybe I accidentally clicked . Probably just user error
< fanquake>
You'll be happy to know that GitHub is flat out rolling out new features while at the same time not addressing our concerns
< jnewbery>
that's fantastic news. Can't wait to update my github story and add some filters to my avatar
< * aj>
clicks like and subscribe just in case
< * fanquake>
will continue to bring up ALL the issues
< aj>
jnewbery: "e" is apparently the keyboard shortcut to mark as done
< luke-jr>
wumpus: is there a guarantee they will finish in order? anyway, someone was telling me batches are limited usefulness because the limit affects them..
< luke-jr>
if it's not an intended property, maybe I will look into fixing it
< wumpus>
luke-jr: I'm fairly sure there is a guarantee that they run (and finish) in order, that's the point of a batch
< wumpus>
the work queue assigns HTTP *requests*, a batch is one HTTP request right?
< wumpus>
I don't understand what could be the problem
< wumpus>
luke-jr: the HTTP worker queue code is in httpserver.cpp, the batch handling code is in rpcserver.cpp JSONRPCExecBatch, it literally iterates over the array of commands, executing them one by one, pushing their result into the result array
< luke-jr>
btw would anyone object to libsecp256k1 commits coming here?
< jnewbery>
\o/\o/\o/
< jonatack>
congrats |o|
< ariard>
yes! congrats
< jamesob>
look at all that extra fuzz. nice!
< sipa>
it's rather fuzzy
< wumpus>
luke-jr: don't think secp256k1 notifications should be here, no, but isn't there a specific channel for commit notifications?
< michaelfolkson>
A new channel for secp256k1 notifications or have notifications set up on the existing #secp256k1 channel?
< michaelfolkson>
I guess a question for the #secp256k1 channel
< luke-jr>
well, I guess there's #bitcoin-commits
< luke-jr>
maybe I just need to make it more visible to myself
< wumpus>
bitcoin-commits did have secp256k1 notifications as far as I know? but don't know if it's still the case since github removed their own IRC notification
< luke-jr>
yes, it does
< wumpus>
ok!
< wumpus>
ah yes it's even the same bot
< michaelfolkson>
Cool, joined it. Thanks
< bitcoin-git>
[bitcoin] sipa opened pull request #20147: Update libsecp256k1 (endomorphism, test improvements) (master...202010_secp256k1) https://github.com/bitcoin/bitcoin/pull/20147
< instagibbs>
argh no email notifications of PRs for me now
< instagibbs>
ie \o/
< andytoshi>
has anyone here implemented BIP322? i am working on an implementation and have some questions about the bip