From 02298d2601b42d12a00cac2ff557bc52b0b4b87a Mon Sep 17 00:00:00 2001 From: AJ Rossini Date: Mon, 18 Jun 2007 08:58:05 +0200 Subject: [PATCH] removed Lisp-specific help --- README.excl | 20 -------------------- README.kcl | 42 ------------------------------------------ README.mcl | 50 -------------------------------------------------- 3 files changed, 112 deletions(-) delete mode 100644 README.excl delete mode 100644 README.kcl delete mode 100644 README.mcl diff --git a/README.excl b/README.excl deleted file mode 100644 index 9031f04..0000000 --- a/README.excl +++ /dev/null @@ -1,20 +0,0 @@ -To build the Allegro version, - - Rename all .lsp files to have .lisp extensions - - Edit Makefile and set - - CFLAGS - EXCL - LSDIR - - do 'make excl' - -After building the executable, copy Examples and Data to ${LSLIB}, -saved_kcls to ${KCLSDIR}, and kcls to an appropriate place on your -search path. - -At this point, I have not firgured out how to tune the memory -management in Allagro. If you find good setting for this, you may want -to add them to defsys.lisp. Also please let me know what they are so I -can add them to the distribution. diff --git a/README.kcl b/README.kcl deleted file mode 100644 index 6bb2ddd..0000000 --- a/README.kcl +++ /dev/null @@ -1,42 +0,0 @@ -There are two ways to build the KCL version. The first, and simpler, -approach is to - - Check defsys.lsp to make sure :stat-float-is-double-floatsure - is NOT pushed onto *features* -- just comment out the - appropriate line - - Edit Makefile and set - - CFLAGS - AKCLDIR - KCKSDIR - LSDIR - - do 'make kcl' - -The second approach requires that you - - Rebuild AKCL using the modified versions of c/num_sfun.c and - lsp/numlob.lsp included here. - - Check defsys.lsp to make sure :stat-float-is-double-floatsure - IS pushed onto *features* -- just comment out the appropriate - line if necessary - - Edit Makefile and set - - CFLAGS - AKCLDIR - KCKSDIR - LSDIR - - do 'make kcl' - -The second approach fixes some questionable choices on the handling of -floating point coercion in AKCL and also fixes asin and acos, which I -think are are redefined incorrectly in ACKL. The second approach -produces code that is about 10% faster on a standard set of tests. - -After building the executable, copy Examples and Data to ${LSLIB}, -saved_kcls to ${KCLSDIR}, and kcls to an appropriate place on your -search path. diff --git a/README.mcl b/README.mcl deleted file mode 100644 index 0801fa0..0000000 --- a/README.mcl +++ /dev/null @@ -1,50 +0,0 @@ -To build Lisp-Stat under Macintosh CL, - - Put the files in a mac folder with the lib, Examples and Data - files in subfolders and rename all .lsp files to have .lisp - endings - - Edit the MPW library specifications in mclglue.lisp if necessary - - Use MPW to build clib.o in the lib folder with the makefile - clib.make - - Start up MCL and from with in MCL - - Load defsys.lisp using the load menu command - - In the listener execute - - (ls::compile-stats) - -This sets up fasl files for all the lisp files. You can then run the system -by - - Starting up MCL - - Loading defsys.lisp with the load menu item - - In the listener executing - - (ls::load-stats) - (ls::use-stats) - -Up to now I have not figured out the right way to write a modified top -level that restores the history mechanism and behaves properly in the -interactive enviroonment. I have also not been able to figure out how -to save an application or image that contains the C code. If anyone -has any hints I would be glad to hear them. Since this part of MCL is -still listed as under development it may be best not to spend too much -time on this until they come out with a proper release. - -The result seems to work but is very unstable on a Mac SE30 with 5M -and system 6.0.7. I don't think the problem is with the C interface -code itself, since it is essentially the same code as for Allegro. I -think the problem is a combination of lack of memory and heap -fragmentation. I have not yet worked out if there ae parameters in mcl -that let you tinker with heap allocation and memory management in -general. If so, these need to be adjusted. More memory would also -help. Finally, it would probably help to write local versions of -calloc/realloc/free that do more heap compacting and garbage -collecting. - -- 2.11.4.GIT