Interlisp.org
Towards ANSI CL - SICL and Clostrophilia #2315
RyanBurnside
started this conversation in
Ideas
Replies: 3 comments 1 reply
-
|
Some time ago @masinter posted about SICL and I seem to remember we had some discussion on it. |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
Perhaps some of the ANSI compliance tests could be adapted to get an exact idea of the current gap. I recall some implementations claiming a percentage of compliance via a common test. There might be caveats for Medley as it's a very unique platform which is understandable if some compliance is not reasonably possible. |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
There are subtleties with CLtL2 transition due to the nickname of the CL package changing from LISP to COMMON-LISP and how this affects files that were written in one package system and have to be read into another. There’s a function FLIP-CL in the COMMON-LISP-PACKAGE source file in the CLTL2 directory that flips between them to allow you to put the system into one namespace, read your files, change the namespace, write the files. The last version of PCL does build in the CLtL2 system, but is missing some stuff in the existing system.
… On Oct 12, 2025, at 4:22 AM, Ryan Burnside ***@***.***> wrote:
I know that bringing the Common Lisp implementation up to speed is a "nice to have" goal.
I realize that the focus is probably getting things as stable as can be first.
On that note, I had a discussion with beach on the #commonlisp IRC channel on irc.libera.chat:6697 and it was suggested that we might consider SICL which aims to be a portable modular project to allow any base dialect of Common Lisp to reach ANSI compliance. There is also Clostrophilia which implements CLOS to spec.
There may be special considerations for Medley but I'm not yet advanced enough to know if the modules in the project would be portable. I do see though that the guys who work on it are VERY experienced Common Lispers.
The following snippit of conversation hints that some CLOS would probably be required.
<Pixel_Outlaw> beach, does SICL aim for ANSI compliance? I had assumed so.
<beach> Yes, SICL is meant to be a conforming Common Lisp implementation.
<Pixel_Outlaw> Ok, thank you very much.
<beach> Sure.
<beach> How is SICL related to Medley?
<Pixel_Outlaw> Well, they've got the bones of CL somewhere between CLTL1 and CLTL2. I'm hoping we can use the modules and pull er into the future. Medley hosts both InterLisp and Common Lisp. You get Xerox Common Lisp and I believe one more modernized varient.
<beach> I see.
<Pixel_Outlaw> Some guessing may have taken place around CLOS. They've got pieces of it and implemented the setf accessors for CLOS but it was before CLOS was finalized.
<beach> Most extracted SICL-related libraries that implement part of the standard require CLOS.
<beach> Oh, OK.
<beach> Clostrophilia contains most of the AMOP specification, except (of course) how objects are represented.
<beach> Like the sections on object initialization, object finalization, etc.
<Pixel_Outlaw> masinter, in the chat there -> was a member of the Common Lisp Standards Committee and represented InterLisp.
<beach> Yes, I know.
<Pixel_Outlaw> Ok.
<Pixel_Outlaw> Well, might just be a matter of seeing if we've got enough CLOS there and the right pieces. Thanks.
<beach> Sure.
<beach> If not, that would be something to work on.
<Pixel_Outlaw> Yes, it would seem a small price to pay for huge gains if modules could be used.
<beach> Oh, and Clostrophilia also contains the published dispatch technique.
<beach> ... but we don't optimize EQL specializers yet.
<Pixel_Outlaw> Ah, so in theroy we might be able to use Xlostrophilia to bring the CLOS up to spec, then continue on with SICL?
<Pixel_Outlaw> *Clostrophilia
<beach> Possibly. I am not sure how that would work out with your bootstrapping technique.
<beach> Something to look into for sure.
<Pixel_Outlaw> Well, it's an exciting slam dunk if there is a possibility!
<beach> Sure.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>
|
Beta Was this translation helpful? Give feedback.
1 reply
-
|
I wonder if with some tweaking we could run the ANSI Compliance Test Suite... |
Beta Was this translation helpful? Give feedback.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I know that bringing the Common Lisp implementation up to speed is a "nice to have" goal.
I realize that the focus is probably getting things as stable as can be first.
On that note, I had a discussion with beach on the #commonlisp IRC channel on irc.libera.chat:6697 and it was suggested that we might consider SICL which aims to be a portable modular project to allow any base dialect of Common Lisp to reach ANSI compliance. There is also Clostrophilia which implements CLOS to spec.
There may be special considerations for Medley but I'm not yet advanced enough to know if the modules in the project would be portable. I do see though that the guys who work on it are VERY experienced Common Lispers.
The following snippit of conversation hints that some CLOS would probably be required.
Beta Was this translation helpful? Give feedback.
All reactions