As string_builder is being renamed to avoid confusion, which is addressed in the stdlib upgrade bd40bea it makes sense to reconsider if we do actually need it.
There are a lot of uses in the generate_bindings module, but it essentially doesn't matter there, as performance is irrelevant, codegen is just used once when generating new bindings for a version of the protocol, generated bindings become part of the package. For the sake of consistency we could just replace it with regular strings, but it's not a priority at all.
There is a use in a hot loop in the chrome actor, in which it serves as a message buffer, here we could attempt to replace it with just a regular string and benchmark the performance difference.
As string_builder is being renamed to avoid confusion, which is addressed in the stdlib upgrade bd40bea it makes sense to reconsider if we do actually need it.
There are a lot of uses in the
generate_bindingsmodule, but it essentially doesn't matter there, as performance is irrelevant, codegen is just used once when generating new bindings for a version of the protocol, generated bindings become part of the package. For the sake of consistency we could just replace it with regular strings, but it's not a priority at all.There is a use in a hot loop in the
chromeactor, in which it serves as a message buffer, here we could attempt to replace it with just a regular string and benchmark the performance difference.