< bitcoin-git>
bitcoin/0.19 4cf7350 Wladimir J. van der Laan: Merge branch '0.19_2' into 0.19
< fanquake>
In case anyone is wondering what has just happened; I had a lapse of concentration while preparing commits for backport (planning on cherry-picking to a branch and opening a PR), and accidentally pushed 9d980c1f7f6fd0eb686d3cc94ba4698547f151c2 straight to the 0.19 branch. wumpus has pushed 4cf73504220e95ea1c4961509ea757ce6d393a9b as a followup to 0.19 to add a Tree-SHA512 hash.
< kallewoof>
Now that you mention it it did look kinda odd. I just assumed it ws because it wasn't master. :)
< wumpus>
we kept it as it was as it's a good backport, avoiding the need to do a force-push
< kallewoof>
FTR, verified the backport matches the original.
< jeremyrubin>
Currently we maintain an entire *data structure* in the mempool just because of a circular resolution issue
< jeremyrubin>
And we then have to do all these extra map lookups too rather than just keeping the data in the owning class
< sipa>
i haven't followed up on the fuzzing work that's been done to the codebase... but i wonder has anyone considered using a dummy signature checker, and fuzzing script execution with it?
< wumpus>
not sure either, would be best to ask practicalswift
< phantomcircuit>
sipa, i dont think anybody has done that
< sipa>
maybe i'll spend some time looking into how fuzzing works :)
< instagibbs>
sipa, you shoudl attend the PR review club they did it 2 weeks ago :)