Free the X3J Thirteen!

A monthly summary of Free ("as in Freedom") CL Hackery news-that-I-know-about. September 2001

For submission information, see the bottom of the page

CLiki

Just in time for this issue of FTX13, a new downloadable release of the CLiki engine was announced. Spooky coincidence, hey?

• Hyperspec lookups
• A full-text search engine
• Multiple CLiki instances can co-exist in the same Lisp instance
• Recent Changes view is now paged for speed of downloading
• Can optionally use a cookie to remember the editors' names

CLISP

A pretest of the upcoming CLISP 2.28 is available. User-visible changes:

• ANSI CL functions MAKE-LOAD-FORM and MAKE-LOAD-FORM-SAVING-SLOTS are implemented.
• The ANSI CL pretty printer (chapter 22.2) is implemented. In particular,
  variables *PRINT-LINES*, *PRINT-MISER-WIDTH*, *PRINT-PPRINT-DISPATCH*,
  macro PPRINT-LOGICAL-BLOCK,
  local macros PPRINT-EXIT-IF-LIST-EXHAUSTED, PPRINT-POP
  and functions PPRINT-DISPATCH, COPY-PPRINT-DISPATCH, SET-PPRINT-DISPATCH, PPRINT-FILL, PPRINT-LINEAR, PPRINT-TABULAR, PPRINT-TAB, PPRINT-NEWLINE, PPRINT-INDENT
  are implemented. See the impnotes for details.
• The functions WRITE and WRITE-TO-STRING now accept :LINES, :MISER-WIDTH and :PRINT-PPRINT-DISPATCH keyword arguments. This change requires recompilation of all .fas files.
• Support Cygwin's "/cygdrive/drive/path" pathname syntax on Win32.
• New macro EXT:FCASE which allows specifying comparison function in CASE.
• Fixed a bug: LDIFF now handles dotted lists correctly.

More information at http://www.geocrawler.com/lists/3/SourceForge/1122/50/6640099/

SBCL users will note that MAKE-LOAD-FORM has always been claimed as the biggest obstacle preventing CLISP from building SBCL. It should get further now, though it still needs more work:

• http://www.geocrawler.com/archives/3/1663/2001/7/0/6289056/
• http://www.geocrawler.com/lists/3/SourceForge/1124/25/6596086/

CMUCL

The new CMUCL machine has arrived and been mostly configured. CVS access for committers is reported to be working again, so peoples' patch queues should be emptying. Anon cvs will be along in a little while.

ECLS

ECLS stands for Embeddable Common-Lisp "Spain", usually read as "e-klos". It is a an implementation of the Common-Lisp language which aims to comply to the ANSI X3J13 definition of the language. ECLS 0.4 has

• Ports to Debian Linux PPC and Cygwin.
• Changes to allow a static build on platforms without dlopen()
• The usual array of bug fixes

See http://ecls.sourceforge.net/news.html for more

Franz

Franz, makers of Allegro CL, have opensourced four new Lisp packages this month:

• Turn your Common Lisp app into a Windows NT/2000 service
• NFS server for MS Windows
• DNS server
• Uniform Resource Indicator (URI) support for Common Lisp

The URI support is apparently the same code as comes with Allegro 6. Given that ACL users already have it, one hopes that Franz will be taking patches to make it portable to other Lisp implementations - otherwise it's going to have to be maintained by someone else (as Portable Allegroserve is) if there's to be any purpose in making it available freely.

(From a brief perusal of the code it looks mostly portable already, though it contains some pre-ANSI constructs and refers to a couple of symbols in the EXCL package. And of course it uses IF*)

More information at http://opensource.franz.com/

GCL, Maxima

Richard Stallman writes on clisp-list

One side consequence of William Schelter's death is that we are
looking for new maintainers for GCL and Maxima.  If anyone wants to
volunteer, please contact me.

OpenMCL

OpenMCL 0.7 was released this month. The new version includes

• A BSD sockets interface
• Much better support for running external programs
• Primitives for manipulating shared libraries
• Support for using AltiVec in LAP
• Primitives for doing Multibyte I/O
• Many changes and fixes to the pathname code
• Documentation on OpenMCL's interface translation system
• Lots of bug fixes.

Clozure Associates have also resolved their ISP troubles and are back on the net.

SBCL

The SBCL SPARC and PPC ports announced last month are both progressing through the "make cold-init work" process. SPARC is presently suffering signal handler problems (SA_SIGINFO doesn't seem to be working as advertised, maybe a newer kernel will help), and PPC is munging its number stack pointer in some odd circumstances to do with multiple value returns

Did the previous paragraph make absolutely no sense? If so, you need to look at the other new SBCL project this month: the SBCL Internals Documentation project -

• http://ww.telent.net/sbcl-internals/index

This is another CLiki-based site wherein SBCL hackers can braindump anything they've found recently and the magic of Wiki will assemble it all into a coherent document. Or failing that, at least the search engine works.

Submissions

"Free The X3J Thirteen" is compiled once a month by me (Daniel Barlow) from news that I know about. If you're associated with a free CL project that's under-represented here, all you have to do is send me news so that I know about it. Email me! I'd especially like to hear news from CLISP and OpenMCL people because I don't often have time to follow that myself