< bitcoin-git>
[bitcoin] kallewoof opened pull request #9235: Refactor: Removes all uses of `using namespace` in all source files. (master...no-using-ns2) https://github.com/bitcoin/bitcoin/pull/9235
< achow101>
what happened to createwitnessaddress?
< btcdrak>
achow101: it was removed
< achow101>
I gathered that much. I don't quite understand why it was removed. Couldn't it just have been modified to check for address validity too?
< sipa>
achow101: it has no access to the wallet, so it would at least need an api change
< sipa>
where you pass it the full script, pubkeys, ...
< achow101>
oh. i see
< fanquake>
Finally finished a deterministic fuzzing cycle with 9172. Happy to share the outputs with anyone who'd like a look.
< btcdrak>
fullnode wrist watches? I like the idea.
< bitcoin-git>
[bitcoin] gmaxwell opened pull request #9236: Fix races for strMiscWarning and fLargeWork*Found, make QT runawayException use GetWarnings (master...strMiscraceless) https://github.com/bitcoin/bitcoin/pull/9236
< Chris_Stewart_5>
When the tx is serialized for a signature, the redeem script has a '43' pushop prepended to it so the script is '4341047...'
< Chris_Stewart_5>
in the case of segwit I should say
< morcos>
sipa: What do you think about removing txConflicted from the block connection logic? You and I had a discussion about it referenced here: #8692 and sdaftuar and I have been discussing all morning
< morcos>
Right now, it does effectively nothing (marks balances dirty, but the balance recalculation will result in the same balance because the txs conflicted via mempool detection aren't actually marked conflicted)
< morcos>
I propose commenting (in the code?) about a general out line of how we could do a better best efforts for also using the mempool for conflict detection. but there are several corner cases that make this a larger project.
< morcos>
I think it woudl be simpler to understand if for now we didn' thave this ineffective method sitting around, and later if anyone is inclined they could try a new design to do it properly?
< morcos>
BlueMatt: ^ ?
< bitcoin-git>
[bitcoin] jonasschnelli opened pull request #9238: Ignore BIP35 mempool command by default (master...2016/11/dis_mempool) https://github.com/bitcoin/bitcoin/pull/9238
< sipa>
morcos: i vaguely remember we're effectively only using the in-wallet confict detection?
< morcos>
sipa: correct, the in-wallet detection is the only one that marks conflicted.. that issue links to the IRC convo you and i had, but in short you were proposing we could make the mempool conflict detection also mark conflicted
< morcos>
and then be smarter later about how to mark things dirty if they might have become unconflicted
< sipa>
hmm
< morcos>
this would improve our "best efforts" basis of finding all conflicts
< morcos>
but i think there are lots of corner cases and its basically impossible to get perfect
< sipa>
my only concern was that it felt stupid to delete code which has a reasonable chance of being close to working
< morcos>
my only suggestion now, is remove the misleading and useless mempool conflict detection for now, until we decide to do it better, if we ever do
< sipa>
but if you say there are many edge cases to get right, delete it
< morcos>
yeah that was my concern too, but i think there are enough issues around it .. yep.. ok, i'll propose a PR
< morcos>
not necessarily too many, but too many that we're about to do it anytime soon, and it might look a bit different when we do
< morcos>
jonasschnelli: I have some questions about the new smart fee slider label introduced in #8989
< morcos>
The way the smart fee slider always worked was it was displaying to you the estimate that estimateSMARTfee was returning which might have been at a different number of blocks than your requested target.
< morcos>
now it is true, that the pre-existing code was then actually trying to send your transaction with estimatesmartefee from you requested target, which i guess might have changed if that was all of a sudden now possible to return an answer
< morcos>
so that was maybe already a slight bug
< morcos>
but i think it is confusing now that the label below the slider shows your requested target, which maybe different from the label above ths slider which shows the found target
< morcos>
That said... i'm not really sure what the solution is
< bitcoin-git>
[bitcoin] morcos opened pull request #9239: Disable fee estimates for 1 block target (master...blockstreamtil2blocks) https://github.com/bitcoin/bitcoin/pull/9239
< morcos>
jonasschnelli: I think the way I did above PR, the problem will be rare that the 2 numbers will be different... so its probably not worth worrying about.. although you should make sure you are ok with changes to smartfeeslider
< Chris_Stewart_5>
Is casting to CScriptBase the only way to serialize CScript to a hex string while keeping the push ops around?
< BlueMatt>
morcos: hum, I dont think I've looked at the conflict-detection code in forever
< BlueMatt>
if you fix it up I'm more than happy to take a look :p
< morcos>
BlueMatt: i'm not fixing it up (at least not now) i'm removing the code that doesn't serve a valuable function.. partly pinging you as it affects #9014. anyway, will do and then we can discuss on PR. right now i'm swimming in asserts
< BlueMatt>
morcos: yea, i figured it'd effect 9014, but probably not in any way I care about :p
< jonasschnelli>
morcos: good point.
< jonasschnelli>
The label does show the requested target
< morcos>
jonasschnelli: once we eliminate fee estimates for 1, they will only disagree either on startup or very rarely.. so i think its probably ok..