Topic: Bitcoin dev IRC meeting in layman's terms (2015-11-19) (Read 421 times)

Once again my attempt to summarize and explain the weekly bitcoin developer meeting in layman's terms. 
Link to last weeks summarization   


Disclaimer

Please bear in mind I'm not a developer and I'd have problems coding "hello world!", so some things might be incorrect or plain wrong. 
Like any other write-up it likely contains personal biases, although I try to stay as neutral as I can. 
There are no decisions being made in these meetings, so if I say "everyone agrees" this means everyone present in the meeting, that's not consensus, but since a fair amount of devs are present it's a good representation. 
The dev IRC and mailinglist are for bitcoin development purposes. If you have not contributed actual code to a bitcoin-implementation, this is probably not the place you want to reach out to. There are many places to discuss things that the developers read, including this forum.


link to this week logs 
Meeting minutes by meetbot 


Main topics discussed where: 
 
transaction priority   
dealing with mempool eviction   
Sequence numbers 


Short topics/notes 

Opt-in replace by fee needs some extra testing, but otherwise seems ready to go. Some wallet developers are on board and actively participating, for example GreenAddress.


transaction priority

- background 

Each transaction is assigned a priority, determined by the age, size, and number of inputs. Which currently makes some transactions free.   
This currently has a large amount of code, which makes it harder to maintain, and is not that optimal since you can't expect miners to include 0-fee transactions. 

- meeting comments

If we don't stop support for priority in transaction creation we also need a mempool area for priority, or those transactions will always get evicted. 
If we develop a better framework to support these kind of metrics we can add it back. 
Plan is to remove the priority transaction creation from the wallet, not the mining part. 


- meeting conclusion 

Creation of priority transactions should be removed from the wallet. 


dealing with mempool eviction

- background 

When a transaction is relayed across the network it is held by the nodes in memory, until it gets into a block. All these transactions that sit in memory are called the memorypool or mempool for short. 
Like we could see during the spam-attack if there's a big back-log of transactions that couldn't make it in the blockchain this mempool can get pretty big resulting in nodes crashing. 

To stop this from happening devs created a mechanism to reject and/or remove transactions from the mempool. 


- meeting comments

Current problem: when a wallet transaction is rejected by the mempool, the wallet considers the resulting transaction as "conflicting" and will happily respend the inputs.   
sipa proposes to make the wallet only treat a transaction as conflicting if it has non-existing inputs. 
It should however consider it respendable at some time later on.   
You could add a way to manually remove transactions, or tag is as removed, or archive it. 
You could also do something separate that marks the transaction as respendable, as removal gives the impression the transaction can't be mined in the future. 
Options that are wanted: a "respend with higher fee" option and an option to forget about a transaction completely, we need a minimum viable idea for 0.12 though. 


- meeting conclusion

Given the tight deadline for 0.12 we detect actual conflict instead of mempool eviction and leave the coins immediately respendable. 

Sequence numbers

- background

BIP 68 repurposes some of the unused nSequence field to a relative locktime, meaning locking inputs until a certain time or blockheight has passed.   


- meeting comments

We need to wait for BIP113 to be deployed as standardness so BIP 68, 112 and 113 can go in a softfork. 
There's upcoming projects that already use sequence numbers. 
Merging BIP68 would make BIP112 easier to review and would stop the need to rebase all the time. 
If we feel the 68/112 are sufficiently reviewed and mature they could go in as standardness rules. 
The BIP text doesn't seem to reflect what's written in the code. 
 

- meeting conclusion

Check BIP68 to match the implementation



Participants