[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index]
[Thread Index]
- Subject: RE: Flushing global variables
- From: "Murray S. Kucherawy" <msk@...>
- Date: Thu, 19 May 2011 23:43:34 -0700
> -----Original Message-----
> From: lua-l-bounces@lists.lua.org [mailto:lua-l-bounces@lists.lua.org] On Behalf Of Tony Finch
> Sent: Thursday, May 19, 2011 2:49 AM
> To: Lua mailing list
> Subject: Re: Flushing global variables
>
> > I'm working on an application that's multi-threaded, and to avoid a
> > newstate-load-pcall-close series for every time I want to invoke a
> > script, I'm thinking of allocating a pool of states and just pulling one
> > out when it's time to run, which leaves me with just load-pcall every
> > time.
>
> Why not cache the parsed scripts, so it's just a pcall each time?
I do, actually; there's a call to lua_load() and then one to lua_dump() to get the compiled form back out. Then when it's time to run the script, I lua_load() the compiled form and then use lua_pcall() to run it. Thus, it only gets parsed once.
> > One thing I'm not certain about with that design: Will the garbage
> > collector purge any global variables that exist in the state on
> > completion of the pcall, or will they linger around for the next
> > load-pcall on the same state object? If the latter, can I use some
> > getfenv/setfenv magic to reset things for the next iteration? Or do I
> > have to bite the bullet and do it the more expensive way?
>
> Yes, it sounds like you want some kind of setfenv sandbox. You can use a
> __index metatable entry on the temporary environment to pull in the
> persistent global functions efficiently.
Thanks, I'll check into this (and the other suggestions).