Re: Ideas about implementing a string type?

[Rici Lake <lua@...>]

David Kastrup wrote:
> One possibility would be that at len=15 (add one byte for
> 0-termination), the length of string variable and content together
> matches a size that makes malloc and/or free do extra work for some
> reason.

There's only one malloc in the test. Take a look at the benchmark
to see what it's measuring.