@yilongl_megaeth@IsaacN0tNewton Hey @yilongl_megaeth I'm having trouble replicating the benchmark, seems to panic when I run it. Log here: https://t.co/V0vNE3Fg7V
I tried to debug it without much success. I'll keep digging.
@united They were not very helpful. I contacted customer service and they said use the customer service app. When I did that, it said I had to call an 800 number. And when I called that, it said they were extremely busy and requested that I call back later.
Always possible to remove unwrap() using compiler logic in #rustlang, and it's a good idea in production code. TIL about `#![deny(clippy::unwrap_used)]` to prohibit it. #firewood
PSA: Don't have two unrelated data structures that are secretly related in code; instead put them into a new structure, lest you create a hodge-podge of structures representing state. #rustlang#firewood
PSA: #rustlang: passing a `mut` to a function is not as good as just adding another return type. Hard: `fn foo(&mut T2) -> T1` vs easy: `fn foo() -> (T1, T2)`
Exception: `Option<&mut T>` is okay if computing T is expensive and the caller sometimes tosses it. #firewood
Early benchmarks for #firewood are showing a huge improvement over the existing database formats used to store blockchain data. Our benchmark numbers are not ready to be published yet as they still include the creation of the keys and such.
whoa, in today's PR comments I wrote: better yet "impl From<&[u8]>" for "NodeType" -- funny how I didn't include "for" in the syntax, since that's the actual #rustlang syntax
Naming is so hard. Just changing the name of a variable from something meaningless to something with meaning makes code so much more readable. #firewood renames happen every day it seems, and it's improving!
Hot from the #firewood discussions with @_richardpringle today, we're talking about renaming shale (https://t.co/hlpuMoynF2), the module that needs a lot of tlc, to woodshed, and doing a full rewrite to reduce copies and deal with lifetime issues.