Forgot to reply to all... Also, re: Luca, as I said in a previous message, it just doesn't work to wholly disable the cache if you don't also disable the ASDF repository. Otherwise you get a lot of systems that you can't load anymore, because they are in the repository, but you have no write access to where they are. And so, even if the cache is controlled by a toggle (instead of an exposed redirection list as I would like), this toggle SHOULD NOT disable cacheing for clc-managed systems. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] Always design a thing by considering it in its next larger context -- a chair in a room, a room in a house, a house in an environment, an environment in a city plan. -- Eliel Saarinen ---------- Forwarded message ---------- From: Faré <fahree@gmail.com> Date: 02-Oct-2006 09:17 Subject: Re: clc: what to force into the cache directories and what not To: Peter Van Eynde <pvaneynd@debian.org> Yes, such systems are broken. But more systems will be. And when you develop, it is useful to have the fasl's in the current directory so that your usual Makefiles work (especially the make clean part). That is why cl-launch has a mechanism for excluding trees from the cache. I propose that clc should have a similar mechanism. Exposing the redirection configuration to the user can only be good. I think such a mechanism should even be pushed into upstream ASDF (after we finalize it), though with an empty configuration by default. That said, we should *also* fix McClim and such. Usually a matter of using (or *load-truename* *compile-file-truename*) instead of *load-truename*. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] We know how syntax is prefix in LISP, whereas FORTH has a postfix syntax and C's syntax is braindeadfix; well, the syntax of XML is a syntax that is a hardcorepornfix syntax: it does it in the all places at the same time. On 02/10/06, Peter Van Eynde <pvaneynd@debian.org> wrote:
Hello,
I'm preparing a clc version that will undo the *redirect-fasl-files-to-cache* 'fix'. As I see it I could claim that the systems that fail when we redirect all fasls to the cache directories are just broken, but that is quite a statement. So what to do?
- claim that slime/McClim is broken and fix them - only redirect to /var/cache for packaged libraries
I'm voting for option 1.
Oh Luca: #384457 seems still closed to me, not?