Re: [Bitcoin-development] Draft BIP for Bloom filtering

[Mike Hearn <mike@plan99.net>]

I'm thinking we should actually make the change we talked about before
and have the filtered block sent before the transaction data.

For one, it's not intuitive (API wise) that you'd get a callback
saying "new pending tx" immediately before another callback saying "tx
was confirmed", but that's what the current setup makes most natural.
To fix it we'd have to notice that a tx message wasn't requested by
us, buffer it, and wait for the corresponding filteredblock message.
It seems cleaner to receive a filteredblock and then for any tx that
matches it, attach it to the FilteredBlock object and wait until it is
full up, then pass it to the wallet code all at once.

Another issue is that to risk analyze unconfirmed transactions you
really have to download all dependencies. That has to be triggered by
seeing an unconfirmed transaction. It's dumb to start this process for
a tx that is actually in the chain, so you need to have some notion of
whether it came from a filtered block anyway. I only realized this
today.

I think when we discussed this before, the justification for having it
work the current way was that it was simpler to integrate with the SPV
client code if it was done this way around. But I don't think it's
really simpler. There are enough odd side effects of doing it this
way, that I feel it'd be better to tweak the protocol now whilst we
have the chance.

On Wed, Jan 16, 2013 at 4:00 PM, Matt Corallo <bitcoin-list@bluematt.me> wrote:
> Actually, there is one more minor algorithmic change I would like to
> make to the way the hash function is computed really quick before it
> gets merged, I'll have that finished up by the end of today.
>
> Matt
>
> On Wed, 2013-01-16 at 11:43 +0100, Mike Hearn wrote:
>> Matts latest code has been tested by Andreas and seems to work
>> correctly. He had to extend the client a bit to refresh the filter
>> every 25k blocks because even with the extra flag, eventually the
>> filter degrades into uselessness, but it did still improve the
>> situation quite a bit.
>>
>> Because it's unit tested, been reviewed by me several times, has an
>> interoperable implementation that has also been tested by Andreas in a
>> build of his smartphone app,  I'm going to ACK the current code and
>> request that it be merged in to 0.8. What do you say Gavin?
>>
>> The next step after that would be profiling. It's a big performance
>> improvement for SPV clients already, but not as much as I anticipated.
>> I suspect there's a simple bottleneck or missed optimization
>> somewhere. But that can obviously come post-0.8
>
>