Performance optimization with slice

[proyb6]

Is it possible to improve performance if the router could convert string to bytes using to_slice?

From the benchmark, to_slice would cost <4 nanoseconds and efficient to compare using ==

Range  17.51M ( 57.11ns) (±10.77%)  32 B/op  14.44× slower
Slice     252.8M (  3.96ns) (± 5.81%)   0 B/op        fastest

Range:
key[0..-2]

Slice:
(key.to_slice)[0,key.size-2]

[stakach]

In the add_route method?
Personally I think key[0..-2] is more readable. It’s not really a method that needs to be optimised either as it is only run during program initialisation

[tbrand]

I can't judge until I see the actual implementation.