Dan Sugalski <[EMAIL PROTECTED]> wrote: > At 12:31 PM +0200 7/14/03, Leopold Toetsch wrote: >>I have moved some interpreter variables (classname_hash, >>compreg_hash, Argv, Env) into one (SArray) list: >>interpreter->iglobals. >>There access isn't time critical, so the extra indirection wont harm.
> This is fine, though many of those things should actually be global > globals rather than interpreter globals, but that's a separate issue. Yep. I think new interpreters should just use the iglobal pointer of the previous interpreter (or pointers to items, depending on what is global). This would need the old interpreter as parameter, when constructing a new one in make_interpreter(). leo
