Antoine Poinsot
Antoine Poinsot
Sigh... Now the MacOS build seems to randomly fail sometimes.
Yeah i'm changing for `emplace_back` and addressing @Sjors' nits right now.
> I have a preference for breaking out the new checks into its own function since if we intend it to someday soon(TM) be a consensus rule it would have...
Extracted the BIP54-specific sigop counting into a separate function that can be moved into `consensus/tx_verify.cpp` with no modification if/when this is implemented in consensus, as suggested by @instagibbs.
Coins are already accessed multiple times in checking an unconfirmed transaction (which should cache them anyways), doing once more shouldn't introduce any noticeable overhead and i like the upside of...
@Sjors `AreInputsStandard` is always called with a warm cache. ~~The benchmark is not.~~ Additional `AccessCoin` calls to fetch coins from the cache do not introduce any meaningful cost. @Sjors @l0rinc...
> The `bench.run` call does not clear `coins`, so it's warm after the first iteration. Actually the benchmark does populate the cache first through `SetupDummyInputs` (otherwise it couldn't work since...
@glozow I agree this needs a ML post. Will take care of this tomorrow. However the motivation for this PR is not to reduce validation times, as it only marginally...
[Posted](https://gnusha.org/pi/bitcoindev/49dyqqkf5NqGlGdinp6SELIoxzE_ONh3UIj6-EB8S804Id5yROq-b1uGK8DUru66eIlWuhb5R3nhRRutwuYjemiuOOBS2FQ4KWDnEh0wLuA=@protonmail.com/T/#u) about this to the mailing list this morning.