stat st_mode wrong for file name = "NUL"
Dead link on FAQ page.
It has already been fixed in git: https://gitlab.com/gnu-clisp/clisp/-/commit/85cb1174de01ea1b7146dba09cc06eae72ebae94 The result has not been published yet, though (waiting for the next release).
CL_MACH_VM configure check will fail with future compilers
Thanks for the report. Fixed through commit 72ae177d.
timestamps beyond 2036 do not work
Thanks for the report! I would not have noticed this in time. Fixed through the commit https://gitlab.com/gnu-clisp/clisp/-/commit/d28822a69d0d6f785764a0f936afbfcb5897c957 .
docs: formatting error in man page: missing newline before macro (with fix)
docs: formatting error in man page: missing newline before macro (with fix)
Thanks for the report. Fixed through https://gitlab.com/gnu-clisp/clisp/-/commit/f2bc9977107f39b0d3fb7478c75b67cecd19bcaf .
princ prints extra newlines
This is a consequence of *PRINT-PRETTY* and *PPRINT-FIRST-NEWLINE* being true. Documented at https://clisp.sourceforge.io/impnotes.html#ppr-first-newline . Not a bug.
lxsession does not update utmp
I agree with at least the first part of your example: NCONC should validate that its arguments except the last one are lists. MAPCAN and MAPCON should check the intermediate results accordingly.
Thank you for the details. Someone else had the same problem on 2019-12-12 and gave this comment: "WARNING: The default installation will wipe out your PATH variable. " I have therefore removed the .exe-format installers from the download directory. Windows users can still use the clisp-2.49-win32-mingw-small.zip and clisp-2.49-win32-mingw-big.zip files. I have also added a note about this bug in the development sources.
Install of CLISP 2.49, deletes Windows 10 Path Statement.
1) Where do you downloaded the CLISP installer from? 2) About your Windows installation: What version is it (32-bit or 64-bit Windows)? 3) Did you see the "deleted" PATH environment variable in the Control Panel? 4) Did you see the "deleted" PATH environment variable in a cmd.exe window? 5) Did you see the "deleted" PATH environment variable through a clisp command (getenv "PATH") ?
Install of CIisp 2.49, deletes entire Windows 10, Path Statement
This is a duplicate of https://sourceforge.net/p/clisp/support-requests/8/.
Install of CLisp 2.49, deletes entire Windows Path Statement
This is a duplicate of https://sourceforge.net/p/clisp/support-requests/8/.
Until someone has reproduced, investigated, and fixed the two issues, your best bet is 1. to use the HEAD of either the master branch or the clisp-2.50 branch at https://gitlab.com/gnu-clisp/clisp , 2. use "make -f Makefile.devel build-porting64-gcc-portability". This is the least optimized variant. You can try out multiple variants through "build-porting64-gcc-portability multibuild-porting64-gcc".
See also issue https://www.cliki.net/Issue%20LOOP-FINALLY-VARIABLES.
make makemake work on macOS
Thanks for the report. Fixed.
CLISP fails ANSI-TEST LOOP.1.40
There are two more related cases: > (loop for i from 1 to 5 finally (return i)) 6 > (loop for i from 1 upto 5 finally (return i)) 6 > (loop for i from 1 below 6 finally (return i)) 6 ANSI CL 6.1.2.1.1 http://www.ai.mit.edu/projects/iiip/doc/CommonLISP/HyperSpec/Body/sec_6-1-2-1-1.html says: "The variable var is bound to the value of form1 in the first iteration and is stepped[1] by the value of form3 in each succeeding iteration, or by 1 if form3 is not provided." Therefore here, the variable I is...
Macro Lambda-Lists Don't Permit early &environment Parameter
When you read section 3.4.1 http://www.ai.mit.edu/projects/iiip/doc/CommonLISP/HyperSpec/Body/sec_3-4-1.html it is clear how the brackets have to be interpreted, namely as an optional part at the given position. (Optional parts at ANY position are written in a different way, see http://www.ai.mit.edu/projects/iiip/doc/CommonLISP/HyperSpec/Body/sec_1-4-1-2-1.html .) Per 3.4.7 http://www.ai.mit.edu/projects/iiip/doc/CommonLISP/HyperSpec/Body/sec_3-4-7.html the &environment var section must come at...
Mark this repository as obsolete.
test failures on mips in socket.tst
Thank you. I have applied your patch.
Thank you for the patch. I will apply it soon. > Maybe a more systematic solution should be designed. Absolutely. There are three options I can currently think of: Extend OS-ERROR so that it includes not only the numeric value of the code, but also a list of known names of this error code. Add a function that converts a numeric errno code to a list of symbolic ones. * Just add a function that maps a symbolic errno code to a numeric one.
Fix possible build failure on some Linux/powerpc machines.
Fix possible build failure on some Linux/powerpc machines.
For this change to work, you need to add the following files from gnulib: m4/iconv.m4 m4/lib-ld.m4 m4/lib-link.m4 m4/lib-prefix.m4 m4/host-cpu-c-abi.m4 m4/asm-underscore.m4 build-aux/config.rpath
Fix memory map related warnings on Alpine Linux 3.7/x86_64.
Fix compilation error in C++ mode with musl libc.
Fix memory map related warnings on Alpine Linux 3.7/i386.
Fix memory map related warnings on Alpine Linux 3.7/x86_64.
Fix compilation error in C++ mode with musl libc.
Fix memory map related warnings on Alpine Linux 3.7/i386.
Fix compilation error in C++ mode (regression from 2018-02-11).
Use newest libffcall.
Fix build failure when building in the srcdir.
Fix build failure on some Linux/mips machines.
Fix build failure on some Linux/i386 machines.
Fix build failure on some Linux/arm machines.
Fix build failure on some Linux/powerpc machines.
Fix possible build failure on some Linux/s390x machines.
Make KERNELVOID32_HEAPCODES work on Linux/arm, Linux/hppa, Linux/m68k.
Bump version number.
Fix possible build failure on some Linux/sparc64 machines.
Fix build failure on some Linux/x86_64 machines (Ubuntu 17.10, Debian).
Fix build failure on some Linux/arm64 machines.
Fix build failure on some Linux/mips64 machines.
Avoid "warning: command substitution: ignored null byte in input".
Comments.
Fix a build failure on some Linux/arm machines.
Improve installation documentation.
Added tag clisp_2_49_92-2018-02-18 for changeset 614b99fdba0c
Fix possible build failure on some Linux/powerpc64 machines.
Bump version number after beta release.
Fix compilation error in C++ mode (regression from 2018-02-11).
Fix possible build failure on some Linux/s390x machines.
Fix build failure when building in the srcdir.
Fix possible build failure on some Linux/sparc64 machines.
Use newest libffcall.
Fix possible build failure on some Linux/powerpc64 machines.
Make KERNELVOID32_HEAPCODES work on Linux/arm, Linux/hppa, Linux/m68k.
Fix build failure on some Linux/mips machines.
Avoid "warning: command substitution: ignored null byte in input".
Fix build failure on some Linux/x86_64 machines (Ubuntu 17.10, Debian).
Comments.
Fix build failure on some Linux/powerpc machines.
Fix a build failure on some Linux/arm machines.
Fix build failure on some Linux/arm machines.
Fix build failure on some Linux/arm64 machines.
Fix build failure on some Linux/i386 machines.
Improve installation documentation.
Prepare a beta release for the Linux distros and testers.
Fix build failure on some Linux/mips64 machines.
option -L russian crashes when no russian locale is present
The attached patch will fix it. Thanks for the report!
linux --with-debug: oint_addr_mask does not cover CODE_ADDRESS_RANGE
Try a build with --enable-portability. This build should not run into this compilation error, and from the lisp.run executable of this build you can try the -mm option.
Please provide the following additional info: grep RANGE config.status Output of ./lisp.run -mm, as described in unix/PLATFORMS.
Implement options -memfile-hash, -memfile-hash-of, -memfile-compatible.
Bump version number.
Implement options -memfile-hash, -memfile-hash-of, -memfile-compatible.
Prepare a beta release for Debian and testers.
Added tag clisp_2_49_90-2018-02-11 for changeset aa240394e736
Simplify code.
Simplify code.
Enable most optimizations on GNU/kFreeBSD/i386.
Add support for GNU/kFreeBSD on i386.
Add support for GNU/kFreeBSD on i386.
Enable most optimizations on GNU/kFreeBSD/i386.
Fix configure output on IRIX.
Enable most optimizations on IRIX mips o32.