< CodeShark>
luke-jr: for BIP123, I think either CC0 or GNU-all-permissive
< sipa>
sdaftuar: well we do have CTransactionRef being a const reference to a CTransaction
< sipa>
sdaftuar: we could have a wrapper class around CTransaction with mutable and self-synchronized fields
< sipa>
that doesn't necessarily contain data that's part of CTransaction itself, but is relevant for validation
< bitcoin-git>
[bitcoin] gmaxwell opened pull request #9594: Send final alert message to older peers after connecting. (master...send_final_alert) https://github.com/bitcoin/bitcoin/pull/9594
< achow101>
gmaxwell: did you broadcast the final alert yet?
< gmaxwell>
achow101: nope. Feel free! but it is set to not relay: so each node gives it out one host at a time.
< achow101>
so the peers who receive it won't broadcast it too?
< gmaxwell>
Correct.
< gmaxwell>
This seemed safer to me in a number of respects.
< gmaxwell>
they'll display it as soon as they connect to are connected from any node running that code and continue to display until they restart.
< achow101>
will you be sending one that does relay?
< bitcoin-git>
[bitcoin] kallewoof opened pull request #9595: [consensus] make coinbase maturity a network parameter instead of an app-wide constant (master...consensus-coin-maturity) https://github.com/bitcoin/bitcoin/pull/9595
< bitcoin-git>
[bitcoin] morcos opened pull request #9596: [bugfix] save feeDelta instead of priorityDelta in DumpMempool (master...dumpmapdeltas) https://github.com/bitcoin/bitcoin/pull/9596
< gmaxwell>
Argh.
< gmaxwell>
"mapDeltas[i.first] = i.second.first;" probably belongs on some "why C++ sucks" page.
< luke-jr>
lol
< luke-jr>
wonder if there's an easy way to make a named-keys pair
< sipa>
use an enum
< sipa>
as key
< sipa>
or a std::tuple with an enum that lists the field positions?
< cfields>
grr, qt's buildsystem is an absolute nightmare
< cfields>
BlueMatt: i know i owe you a PR. Just now coming up for air after battling qt all day
< fanquake>
cfields need a hand with anything?
< bitcoin-git>
[bitcoin] practicalswift closed pull request #9590: Improve readability by removing redundant casts to same type (master...remove-redundant-casts) https://github.com/bitcoin/bitcoin/pull/9590
< bitcoin-git>
[bitcoin] kallewoof opened pull request #9597: [rpc] Bug-fix: getdifficulty: use network parameters for minimum difficulty (master...getdifficulty-use-net-params) https://github.com/bitcoin/bitcoin/pull/9597
< jonasschnelli>
IMO this assumption is not really stable or even broken
< jonasschnelli>
Haven't verified it, but the CKeyPool item at begin of the set must not always be the one with the olders nTime
< jonasschnelli>
*oldest
< jonasschnelli>
especially when using ReturnKey
< bitcoin-git>
[bitcoin] kallewoof closed pull request #9597: [rpc] Bug-fix: getdifficulty: use network parameters for minimum difficulty (master...getdifficulty-use-net-params) https://github.com/bitcoin/bitcoin/pull/9597
< bitcoin-git>
[bitcoin] practicalswift opened pull request #9598: Improve readability by removing redundant casts to same type (on all platforms) (master...remove-redundant-casts) https://github.com/bitcoin/bitcoin/pull/9598
< bitcoin-git>
[bitcoin] kallewoof reopened pull request #9597: [rpc] Bug-fix: getdifficulty: use network parameters for minimum difficulty (master...getdifficulty-use-net-params) https://github.com/bitcoin/bitcoin/pull/9597
< jonasschnelli>
Re: GetOldestKeyPoolTime. The assumption that the first item in the setKeyPool is always the oldest generated key holds.
< bitcoin-git>
[bitcoin] rebroad opened pull request #9599: [WIP] Perform validation in a separate thread to message handling. (master...ValidationThread) https://github.com/bitcoin/bitcoin/pull/9599
< btcdrak>
BlueMatt: I thought we agreed it would make it for 0.14
< BlueMatt>
btcdrak: too late, didnt get merged and we froze
< cfields>
sdaftuar: for the bip68-sequence test, i'm confused about how you tried to fix. Did you also try switching back to GetTime() rather than doubling down on GetTimeMicros()?
< cfields>
sdaftuar: or were you just trying to nail down what actually depends on the mocktime?
< sdaftuar>
cfields: i only tried to double-down on GetTimeMicros(), figuring that otherwise a mocktime that sets time too far into the future would trigger a bunch of disconnects?
< cfields>
sdaftuar: i think there's a hack to account for that, sec
< sdaftuar>
so maybe it would work to just go back to GetTime for everything? or add a way for mocktime to be used with GetTimeMicros, which I think some people have suggested before
< cfields>
well, i think there are a few things we should do
< cfields>
for one (though i'm not sure how much sense this makes practically), i'd like to see any consensus-critical time checks done outside of net code
< cfields>
so that we can at least separate the things we're testing
< cfields>
second, net code needs to move to a steady_clock, so we need to be mindful of how that conflicts with a mocktime
< cfields>
(i don't think that the message timestamp is actually used in validation, but it apparently comes into play)
< sdaftuar>
it gets passed in a couple places in net_processing, but I don't think it goes anywhere except for ping/pong handling?
< cfields>
ok, yep. good. I was afraid that the mocktime was trickling in from there, but apparently not.
< cfields>
sdaftuar: so maybe take the changes you tried, as well as killing the time reset in the setmocktime rpc?
< sdaftuar>
cfields: still failing, though i haven't dug in yet. stepping away but i'll investigate more this afternoon
< bitcoin-git>
[bitcoin] morcos opened pull request #9602: Remove coin age priority and free transactions - implementation (master...nopriority) https://github.com/bitcoin/bitcoin/pull/9602
< bitcoin-git>
[bitcoin] isle2983 opened pull request #9603: Add basic_style.py to automate some style checking. (master...PR-basic-style) https://github.com/bitcoin/bitcoin/pull/9603
< bitcoin-git>
[bitcoin] jnewbery opened pull request #9604: [Trivial] add comment about setting peer as HB peer. (master...HBpeercomment) https://github.com/bitcoin/bitcoin/pull/9604
< bitcoin-git>
[bitcoin] TheBlueMatt opened pull request #9605: Use CScheduler for wallet flushing, remove ThreadFlushWalletDB (master...2017-01-cscheduler-cleanups) https://github.com/bitcoin/bitcoin/pull/9605
< sdaftuar>
cfields: looks like fixing the setmocktime rpc thing + replacing one GetTime() -> GetTimeMicros call that I missed is enough to fix things up, will open a PR shortly
< cfields>
sdaftuar: great, thanks.
< cfields>
meantime, i've coded up a steady_clock implementation that takes system time out of the equation, but it's invasive and will take a while to get in i'm sure
< sdaftuar>
oh cool
< bitcoin-git>
[bitcoin] sdaftuar opened pull request #9606: net: Consistently use GetTimeMicros() for inactivity checks (master...2017-01-net-time-comparisons) https://github.com/bitcoin/bitcoin/pull/9606
< bitcoin-git>
[bitcoin] practicalswift opened pull request #9607: Remove redundant semicolons in Python code (master...remove-redundant-semicolons-in-python-code) https://github.com/bitcoin/bitcoin/pull/9607