< wumpus>
^^ should I create a new issue for the git-subtreecheck issue? is it expected behavior or not?
< instagibbs_>
not to continue off-topic too long, but looks like the consolidations are tether-related or something, it's a *ton* of near dust outputs being put together into 0.0007 BTC chunks
< instagibbs_>
546 sat outputs, all of them
< BlueMatt>
grrrrrr, intel errata, apparently new intel cpus can misexecute jcc if it lands on a 32 byte boundary......
< fanquake>
BlueMatt have you got the new slow microcode yet
< BlueMatt>
fanquake: I'm mostly not on intel these days :p
< BlueMatt>
I just dont, you know, want to fall out of consensus with all the buggy intel cpus
< fanquake>
BlueMatt heh. I guess there isn't Intel chips in those LORA devices either
< BlueMatt>
lol theres no "regular" cpu at all if you do it right
< BlueMatt>
ohoh, i misread that, so the microcode fixes it, just the compiler can fix a bit of the performance loss
< BlueMatt>
thats fine then
< luke-jr>
wumpus: x.y.z - z is bugfixes- totally makes sense to increment it; y is features, not so logical in this case
< luke-jr>
BlueMatt: maybe we should check that the user doesn't have a broken microcode, though
< bitcoin-git>
[bitcoin] jamesob opened pull request #17487: coins: add `erase` parameter to control cacheCoins drop on flush (master...2019-11-au-coins-erase) https://github.com/bitcoin/bitcoin/pull/17487
< BlueMatt>
luke-jr: is there an easy way to do that? Also, we've never done that in the past for intel stupidity, so dunno if we need to in like a minor bump, would maybe be a cool feature in the future, though
< BlueMatt>
given we already have all the code to check cpu flag shit
< sipa>
is there even a way for user level code to gain information about microcode versions?
< luke-jr>
well, we could always explicitly test for the bug?
< luke-jr>
get a 32-byte-aligned buffer and execute jcc?
< luke-jr>
seems like a fit for the current sanity test stuff
< BlueMatt>
luke-jr: according to the errata its some edge case depending on exact internal state of the various icaches and ucode translation shit
< luke-jr>
bleh
< sipa>
if it was trivially observable this bug wouldn't have existed for years
< sipa>
so i think we can assume that it's extremely unlikely to be triggered by chance
< luke-jr>
did it? I intepreted BlueMatt's "new intel cpus" to mean stuff they just started shipping :x
< bitcoin-git>
bitcoin/master 5e8a563 Sebastian Falbesoner: test: add unit test for non-standard txs with too large scriptSig
< bitcoin-git>
bitcoin/master f92e750 MarcoFalke: Merge #17480: test: add unit test for non-standard txs with too large scri...
< bitcoin-git>
[bitcoin] MarcoFalke merged pull request #17480: test: add unit test for non-standard txs with too large scriptSig (master...20191114-test-check-for-non-standard-txs-with-too-large-scriptsig) https://github.com/bitcoin/bitcoin/pull/17480
< instagibbs_>
achow101, I'm not 100% sure that effective value knapsack is going to end up with the "same" results as non-effective value
< instagibbs_>
maybe a bug, but also could just be different behavior coming out of it