IRC meeting summary for 2017-01-05
Notes / short topics
- Jonasschnelli created a draft video for the 2016 Git visualization.
- fork warning system repair
- 0.14 review priorities
fork warning system repair
The large-work fork warning system, which should warn users if a fork with enough proof-of-work continues to be mined, is currently broken under header first validation.
Jl2012 is working on a fix for this.
Jl2012 elaborates this is more than a fix, in that it stores valid headers under an invalid block as long as there’s valid PoW and valid nTime.
Gmaxwell has doubts about the utility of the warning system: there’s been too many false warnings and it wasn’t used usefully by most to begin with. BlueMatt notes having a reliable warning system is the first step towards fixing the trust in the system.
Sipa wonders if we need a detection of internal consensus errors like database corruption, CPU overheating, etc., because 99.99% of all fork warnings that are seen in practice are just broken nodes.
No one is making the argument for this change to be important for 0.14
- Focus more on the 0.14 critical changes
0.14 review priorities
Things people would like to see in 0.14 if review effort wasn’t a bottleneck:
Wumpus would like to see named arguments.
BlueMatt thinks parallel ProcessMessages would be a huge improvement for some use cases, but if there’s not enough time to review it he’ll skip it for 0.14.
Morcos would like to see the block relay improvements (#9375, #9441 and #9447) and possibly parallel ProcessMessages going in. #9441 has a lot of commits, but most of them are very tiny to make review easier.
Gmaxwell was hoping to see the multiwallet support go in (PR’s #8775 and #8694) and the UTXO scriptpubkey index, although the latter looks to be abandoned by its requester. Also the CreateTransaction change from Morcos which fixes some concerning fee overpayment corner case would be nice.
Jonasschnelli thinks HD chain split should also go into 0.14, plus some form of HD rescan.
There’s already a lot of stuff to do, and there’s at least 4 regressions which are 0.14-blocking (issues #9479, #9027, #9148 and #9212). Morcos notes if #9371 (fixing issue #9479) won’t make it in 0.14 we need to revert #9240.
- Focus on fee fixes (#9404), net locks overhaul (#9441), named arguments (#8811),early compact block relay (#9375, #9441 and #9447) and in lower priority the multiwallet changes (#8775 and #8694).
|wumpus||Wladimir van der Laan|
This summary was compiled without input from any of the participants in the discussion, so any errors are the fault of the summary author and not the discussion participants.