Jean-Yves Migeon wrote: > Le 07/10/2013 12:05, Alan Barrett a écrit : > >I still haven't seen a use case for in-kernel Lua. I mean, an > >example (preferably a working example) of something useful that could > >not easily be done without in-kernel Lua. I'd prefer not to see it > >added to the base system without a use case. > > I second the use case. > > Not something as polished or finished as possible, but at least > shows that it is useful (I am well aware of the cause/effect circle > ie. you cannot prove it without having Lua first available, but > breaking that vicious circle with a few examples can help).
In the early days of bpfjit when I didn't yet know of sljit, I was considering ripping off lua code for generating machine instructions from LuaJIT2 code. I still believe that rewriting bpfjit in Lua would improve readability. I even started a rewrite mostly as a good use-case for Lua bindings for sljit but it's low priority project for me. Alex