On Fri, Nov 7, 2014 at 11:09 AM, Tudor Bosman <tudorb@xxxxxxxxx> wrote: > Couldn't LuaJIT, for example, do full (dense) hashing for strings of length > of less than 100 or so bytes, and assume that all longer strings are data > and therefore likely to be unique (and intern them without any hashing / > comparison check)? Is there an assumption that two strings are equal if *and > only if* their interned ids are the same? doesn't plain Lua do something like that? -- Javier