Re: space leak in redefinitions
Jonathan A Rees <[email protected]> Mon, 7 Mar 2011 11:33:55 -0500
| Newsgroups | gmane.lisp.scheme.scheme48 |
|---|---|
| Message-ID | <[email protected]> |
How will out-of-process compilation work (eg the tethered robot system)? The architecture ought to always move toward deeper isolation between compile time and run time, making this a synchronization issue, not a unification issue. This is hard, so I'll be interested to see how you deal with this in a redesign. Jonathan -- apologies for brevity / using handheld gizmo -- On Mar 7, 2011, at 2:56, Michael Sperber <[email protected]> wrote: > > Taylor R Campbell <[email protected]> writes: > >> The definition of CLOBBER-BINDING! in bcomp/binding.scm doesn't remove >> the binding's location if a variable is being redefined as syntax. >> This can leak space, and leads to a weird state where (v ...) refers >> to the syntax and v outside an operator position refers to the old >> variable. > > Good point. The different treatment of value and syntax bindings is a > pain in general. The new macro expander I've been working on doesn't > have this problem anymore, so it'll go away once that's done. > > -- > Cheers =8-} Mike > Friede, Völkerverständigung und überhaupt blabla >