[WIP] More Advanced Transaction Filtering #389
Conversation
CapCap
force-pushed
the
transaction_filters
branch
2 times, most recently
from
cb2d072 to
30ec90f
Compare
CapCap
force-pushed
the
transaction_filters
branch
4 times, most recently
from
52860cd to
282205f
Compare
| WriteSetChangeFilter(WriteSetChangeFilter), | ||
| } | ||
|
|
||
| impl Filterable<Transaction> for APIFilter { |
There was a problem hiding this comment.
You could use enum_dispatch / ambassador to help eliminate some of the boilerplate here.
| pub trait Filterable<T> { | ||
| /// Whether this filter is correctly configured/initialized | ||
| /// Any call to `is_valid` is responsible for recursively checking the validity of any nested filters | ||
| fn is_valid(&self) -> Result<(), Error>; |
There was a problem hiding this comment.
I imagine the reason we have is_valid is because we can't do it in constructors because these structs might get built directly from configuration, and we can't do it in the parsing logic because serde doesn't provide powerful enough attributes to let us do validity checking at parsing time, yeah?
There was a problem hiding this comment.
in the constructors of what?
My assumption for these filters is they will be around for a long time, and different refs will flow through them- there's only ever one copy of these
Injecting it into the serialization would be on the GRPC side, which would be inside of poem, and I wouldn't want to mess with that 😬
| } | ||
| } | ||
|
|
||
| fn is_allowed(&self, txn: &Transaction) -> bool { |
There was a problem hiding this comment.
I don't wanna nerd snipe you but it'd be neat if this function at the top level told you which filter allowed it if this is true 😛
There was a problem hiding this comment.
Like if a transaction got filtered out, it'd be cool to see which filter exactly it failed on.
There was a problem hiding this comment.
sigh :-p
this is not technically hard, just a pain in the ass hahaha... Follow up PR?
CapCap
changed the title
CapCap
force-pushed
the
transaction_filters
branch
from
282205f to
1492324
Compare
CapCap
force-pushed
the
transaction_filters
branch
from
1492324 to
1891c31
Compare
CapCap
force-pushed
the
transaction_filters
branch
from
e3ede30 to
a74334b
Compare
an MVP version of the transaction filtering crate
Performance
On a random 3mb proto of 1000txns during taptos, running the "everything" filter (looped) gives:
Random 10mb proto from April 25th with "everything" filter:
Graffio 100mb proto ("everything" filter, but with
or):