On Fri, Jul 17, 2009 at 3:04 PM, Ben Morrow <ben@morrow.me.uk> wrote: In the end, I find the idea of having to create an object with a DESTROY > method when all you want to do is run some code at the appropriate time > rather messy. Me too. And supporting the "DESTROY as scope-end hack" also wrecks the possibility of switching garbage collection paradigms. (how much mindshare did the proposed "volatile" attribute get? Probably not a lot, since in the absence of alternate GC methods everything is volatile anyway, rendering C<:volatile> a no-op) One further issue here that nothing currently addresses is that things > that are supposed to have lexical scope shouldn't leak out into do FILE/ > require/use and should leak out into eval STRING. "Lexical" may be the wrong word then. "Local" semantics are closer, for when you want to set something that might not "leak out" but is /visible/ from do-file. (several paragraphs thinking out loud about trying to change C<local> to use aliasing instead of assignment removed)Thread Previous | Thread Next