< jonasschnelli>
same result from a different host
< jonasschnelli>
I guess your facing the same issue I recently had.
< sipa>
hmm
< sipa>
and with x9?
< jonasschnelli>
empty response
< jonasschnelli>
I guess it could be a bug in my filter PR
< jonasschnelli>
something with the cache
< sipa>
i already fixed a bug in it
< sipa>
maybe i made it worse :)
< jonasschnelli>
heh...
< jonasschnelli>
Do you take a look at it or should I?
< sipa>
i'm about to fall asleep :)
< rabidus_>
where are you from? if i may ask
< sipa>
i'm from belgium originally, but i'm on the us west coast now
< jonasschnelli>
Yes. Have your rest,.. maybe restart your seeder if you have a chance
< jonasschnelli>
Will try to track it down
< sipa>
ok, restarted
< jonasschnelli>
I have two connected peers, peer A creates a WTX and inv's that to peer B (peer B throws it away [my local change]), I stop both peers [=empty mempools], start again both peers, peer A relays WTX again, but I can't find a inv to peer B (= mempools out of sync, peer A has WTX while peer B don't)
< jonasschnelli>
I guess the "known inventory" will not be persisted on disk? Why does peer A not inv the WTX again to peer B after restart/sweeping the mempool?
<@wumpus>
transactions are only inved when they get relayed, not at any later time, there is no re-inv for old transactions (except by the wallet)
< MarcoFalke>
What is the reason for AGPL in #9049
< gribble>
https://github.com/bitcoin/bitcoin/issues/9049 | Remove duplicatble duplicate-input check from CheckTransaction by TheBlueMatt · Pull Request #9049 · bitcoin/bitcoin · GitHub
< MarcoFalke>
Has someone accidentally uploaded the MongoDB source code to the blockchain, block 413567?
<@wumpus>
what
<@wumpus>
I think he's trolling, but it can't be accepted in this state
<@wumpus>
let's leave the license drama to other altcoins
< jonasschnelli>
wumpus: but when I start a node and the wallet re-broadcasts a tx, it should get inved to the connected peers? Or does the connecting happens after the wallet-re-broadcast?
<@wumpus>
jonasschnelli: yes, if the wallet re-broadcasts it should be inved to connected peers
<@wumpus>
(unless it is known that they know of the transaction, but this information is cleared on restart/reconnect)
<@wumpus>
if connecting happens after rebroadcast yea you'd have to wait until the next one
<@wumpus>
(or rebroadcast it manually using sendrawtransaction)
< jonasschnelli>
Not sure if it's problematic,... just stumbled over this while writing an RPC test
<@wumpus>
jonasschnelli: IIRC that isn't supposed to happen in ReacceptWalletTransactions, that just introduces them into the mempool
<@wumpus>
there's separate logic for the rebroadcast, which doesn't do it immediately at start but periodically, this preserves a little bit of privacy
< jonasschnelli>
hmm... need to track it down further... just detected that RPC sync_all() does not succeed after re-starting two nodes, the mempool wtx from node A does not get over to node B.
<@wumpus>
it shouldn't: there is no automatic mempool sync on start, the sync is just a wait for current transactions to be relayed, if I recall correctly
<@wumpus>
it indeed doesn't handle the scenario where one of the nodes has shut down, that would need additional support
< jonasschnelli>
Okay. I see.
<@wumpus>
I don't think there is any logic, either inthe tests or elsewhere, for doing a node-to-node mempool sync
<@wumpus>
although if the mempool is saved to disk now that'd be possible through that route
< luke-jr>
FYI Knots used to have such logic, but it didn't work well, so I removed it
< luke-jr>
(I'm pretty sure it delayed the real sync until download of peer mempools completed)
<@wumpus>
yes it at most should be something specifically called in the tests, not used in all circumstances
<@wumpus>
(and certainly not use the P2P mempool command)
< jonasschnelli>
A script that truns the milli-timestamps in debug.log to micro-timestamp... meh
< jonasschnelli>
*turns
<@wumpus>
it's clear that there are not enough time formats for logging, what about number of Planck time units since the big bang
< jonasschnelli>
heh
<@wumpus>
it'd even fit into an arith_uint256
< GitHub123>
[bitcoin] paveljanik opened pull request #9063: Do not shadow variable, use deprecated MAP_ANON if MAP_ANONYMOUS is not defined (master...20161102_fix_lockedpool) https://github.com/bitcoin/bitcoin/pull/9063
< BlueMatt>
wumpus: sorry, the license there is the FIBRE license, from which i copied that file
< GitHub68>
bitcoin/master 4a9f3c5 Pavel Janík: Do not shadow variable, use deprecated MAP_ANON if MAP_ANONYMOUS is not defined.
< GitHub68>
bitcoin/master f2a6e82 Wladimir J. van der Laan: Merge #9063: Do not shadow variable, use deprecated MAP_ANON if MAP_ANONYMOUS is not defined...
< GitHub77>
[bitcoin] laanwj closed pull request #9063: Do not shadow variable, use deprecated MAP_ANON if MAP_ANONYMOUS is not defined (master...20161102_fix_lockedpool) https://github.com/bitcoin/bitcoin/pull/9063
<@wumpus>
BlueMatt: it was a good troll :) I'm happy MarcoFalke was awake
< BlueMatt>
yes, wouldve been hilarious if that ended up in bitcoin
< GitHub86>
bitcoin/master 159597a isle2983: [devtools] script support for managing source file copyright headers...
< GitHub86>
bitcoin/master bae81b8 Wladimir J. van der Laan: Merge #8674: tools for analyzing, updating and adding copyright headers in source files...
< GitHub54>
[bitcoin] laanwj closed pull request #8674: tools for analyzing, updating and adding copyright headers in source files (master...copyright-scripts) https://github.com/bitcoin/bitcoin/pull/8674
< BlueMatt>
heh, I hadnt looked at how much better fibre is than the old relay network.... https://imgur.com/a/CpRk8 top one is fibre, note that the top line (all nodes) is equivalent to the graph for rn (bottom), but it never reaches 200ms, whereas the dot on the rn graph is about 200ms
< BlueMatt>
graph is percentile -> time-to-reach-global-coverage
< BlueMatt>
log scale
< achow101>
gmaxwell: I just fired up a 0.12.0 node and I don't see any warning about it being obsolete
< instagibbs>
BlueMatt, impressive
< GitHub126>
[bitcoin] s-matthew-english opened pull request #9064: unify capitalization of "bitcoin address" (master...changeCaps) https://github.com/bitcoin/bitcoin/pull/9064
< achow101>
gmaxwell: alert time?
< gmaxwell>
achow101: about to send.
< gmaxwell>
achow101: it's sent.
< jtimon>
can I run unittests individually ?
< BlueMatt>
i think so
< BlueMatt>
yes, you want test_bitcoin --run_test=
< BlueMatt>
see-also test_bitcoin --help
<@wumpus>
yes, this is documented in src/test/README.md
< GitHub191>
[bitcoin] laanwj opened pull request #9065: Merge `doc/unit-tests.md` into `src/test/README.md` (master...2016_11_unit_tests_doc) https://github.com/bitcoin/bitcoin/pull/9065
< michagogo>
The plan is to send a final-1 alert, wait a while, send a final alert, then post the key, right?
< achow101>
michagogo: yes
< michagogo>
Why not just send the pre-final one, and let the rest sort itself out? :-P
< achow101>
the timeframe is something like over the course of 6 months
< achow101>
the final alert is the max sequence alert which locks out any possible other alert
< michagogo>
achow101: right
< michagogo>
i.e. once the final-1 alert is published, there can't be any other alert but the final one
< michagogo>
Which is hard-coded, right?
< achow101>
right
< achow101>
the final is just to thoroughly kill it off
< michagogo>
Well, it's only *really* necessary to send the pre-final, then publish the key a while later
< achow101>
right
< michagogo>
(And the final one will surely sort itself out :-P )
< GitHub166>
[bitcoin] domob1812 opened pull request #9066: trivial: Remove variable unused after refactoring (master...unused-variable) https://github.com/bitcoin/bitcoin/pull/9066
< gmaxwell>
BlueMatt: so with no longer disconnecting on invalid compact blocks, we need to reintroduce a mechenism so that we kick peers that are on a invalid-to-us forked chain.
< BlueMatt>
gmaxwell: thats still there - we still dos on AcceptBlockHeader results
< gmaxwell>
BlueMatt: e.g. say we have 8 peers, and all of them are feeding us invalid blocks because they're all accepting a chain which is invalid to us. We need to kick some of them so we find peers that are on a chain we accept.
< GitHub151>
[bitcoin] jnewbery opened pull request #9069: Clean up bctest.py and bitcoin-util-test.py (master...btutiltestcleanup) https://github.com/bitcoin/bitcoin/pull/9069