ANSI Common Lisp is a high-level, general-purpose programming language.
GNU CLISP is a Common Lisp implementation by Bruno Haible of Karlsruhe
University and Michael Stoll of Munich University, both in Germany.
It mostly supports the Lisp described in the ANSI Common Lisp standard.
It runs on most GNU and Unix systems (GNU/Linux, FreeBSD, NetBSD, OpenBSD,
Solaris, Tru64, HP-UX, BeOS, NeXTstep, IRIX, AIX and others) and on
other systems (Windows NT/2000/XP, Windows 95/98/ME) and needs only
4 MB of RAM.
It is Free Software and may be distributed under the terms of GNU GPL,
while it is possible to distribute commercial proprietary applications
compiled with GNU CLISP.
The user interface comes in English, German, French, Spanish, Dutch,
Russian and Danish, and can be changed at run time.
GNU CLISP includes an interpreter, a compiler, a debugger, CLOS, MOP,
a foreign language interface, sockets, i18n, fast bignums and more.
An X11 interface is available through CLX, Garnet, CLUE/CLIO.
GNU CLISP runs Maxima, ACL2 and many other Common Lisp packages.
More information at
<http://clisp.cons.org/>,
<http://www.clisp.org/>,
<http://www.gnu.org/software/clisp/> and
<http://clisp.sourceforge.net/>.
Sources and selected binaries are available by anonymous ftp from
<ftp://ftp.gnu.org/pub/gnu/clisp/>
and its mirrors.
2.39 (2006-07-16)
=================
User visible changes
--------------------
* SAVEINITMEM now accepts :SCRIPT argument that disables interpreting
the first positional argument as the script name; and :DOCUMENTATION
argument that is printed by the new -help-image command line option.
See <http://clisp.cons.org/impnotes/image.html> and
<http://clisp.cons.org/impnotes/clisp.html#opt-help-image> for details.
* FFI:UINT64 and FFI:SINT64 are now compatible with C's long long type.
* Stack overflow detection and recovery finally work fine on Unix.
Libsigsegv is required for this, on all platforms (including MS-Windows).
CLISP should neither exit nor crash under infinite recursion.
If your distribution has CLISP compiled without libsigsegv, report
the missing feature to its maintainer.
Note that libsigsegv 2.4 is required, there are bugs in libsigsegv 2.3!
* It is now possible to specify the default method-combination of a generic
function, to be used when the DEFGENERIC form does not specify the
:METHOD-COMBINATION explicitly, through a default initarg specification
for the :METHOD-COMBINATION keyword on the generic function class.
* Readline completion works with non 1:1 terminal encodings, e.g. UTF-8.
* WITH-KEYBOARD works with a Unix tty even when SLIME hijacks *TERMINAL-IO*.
* I/O operations on Win32 are now much faster.
* New functions: POSIX:FFS, POSIX:PATHCONF.
* Infrastructure:
+ Top-level configure now accepts a new option --with-gmalloc to use the
GNU malloc implementation instead of the one supplied by libc.
You may need it on older HP-UX and newer OpenBSD systems.
See file unix/PLATFORMS for more information.
+ The value of the environment variable CFLAGS is respected by configure.
* Bug fixes:
+ SOCKET:SOCKET-SERVER :INTERFACE now behaves as documented.
+ EXT:READ-BYTE-NO-HANG and SOCKET:SOCKET-STATUS used to hang on
buffered binary sockets.
+ Allow DESTRUCTURING-BIND (a . b) with circular and dotted lists.
+ ADJUST-ARRAY of zero length adjustable string now works.
+ TIME now reports correct results when the heap grows over 4GB.
+ RAWSOCK functions now handle :START/:END arguments correctly.
+ BDB:DBC-GET now accepts :READ-COMMITTED and :READ-UNCOMMITTED.
+ POSIX:GROUP-INFO and POSIX:USER-INFO now handle errors correctly.
* Portability:
+ Support DragonFly BSD.
--
Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 5 (Bordeaux)
http://memri.org http://mideasttruth.com http://camera.org
http://iris.org.il http://palestinefacts.org http://jihadwatch.org
Please wait, MS Windows are preparing the blue screen of death.
Javier wrote:
> Thank you.
>
> When is CLISP going to support Intel Mac OSX?
I happened to compile CLISP from CVS yesteday on Intel OSX without any
problems. I couldn't get libsigsev to work right away, so I forced
CLISP to ignore it.
CLISP appeared to install and work correctly, I could run the Slime
backend and evaluate forms.
Cheers
Brad
bradb wrote:
> Javier wrote:
> > Thank you.
> >
> > When is CLISP going to support Intel Mac OSX?
>
> I happened to compile CLISP from CVS yesteday on Intel OSX without any
> problems. I couldn't get libsigsev to work right away, so I forced
> CLISP to ignore it.
> CLISP appeared to install and work correctly, I could run the Slime
> backend and evaluate forms.
Sure, but without libsigsev there is not generation GC (that is said in
the INSTALL file).
Javier wrote:
> bradb wrote:
> > Javier wrote:
> > > Thank you.
> > >
> > > When is CLISP going to support Intel Mac OSX?
> >
> > I happened to compile CLISP from CVS yesteday on Intel OSX without any
> > problems. I couldn't get libsigsev to work right away, so I forced
> > CLISP to ignore it.
> > CLISP appeared to install and work correctly, I could run the Slime
> > backend and evaluate forms.
>
> Sure, but without libsigsev there is not generation GC (that is said in
> the INSTALL file).
Does that mean there is no alternative GC? I noticed the warnings upon
install about no generational GC, but I would expect if there was no GC
at all I would get big screaming letters telling me that CLISP was
cripped by lack of a GC.
libsigsegv appears to have worked on Darwin i686 as recently as 2.4, I
can't imagine that there are huge changes/bugs preventing libsegsegv
from working on OSX Intel.
Cheers
Brad
bradb wrote:
> Javier wrote:
>> bradb wrote:
>>> Javier wrote:
>>>> Thank you.
>>>>
>>>> When is CLISP going to support Intel Mac OSX?
>>> I happened to compile CLISP from CVS yesteday on Intel OSX without any
>>> problems. I couldn't get libsigsev to work right away, so I forced
>>> CLISP to ignore it.
>>> CLISP appeared to install and work correctly, I could run the Slime
>>> backend and evaluate forms.
>> Sure, but without libsigsev there is not generation GC (that is said in
>> the INSTALL file).
> Does that mean there is no alternative GC? I noticed the warnings upon
> install about no generational GC, but I would expect if there was no GC
> at all I would get big screaming letters telling me that CLISP was
> cripped by lack of a GC.
> libsigsegv appears to have worked on Darwin i686 as recently as 2.4, I
> can't imagine that there are huge changes/bugs preventing libsegsegv
> from working on OSX Intel.
I actually had problems installing libsigsev on a OS X PowerPC system,
so I always install clisp without it. (Don't remember the details,
though...)
Pascal
--
My website: http://p-cos.net
Closer to MOP & ContextL:
http://common-lisp.net/project/closer/
Pascal Costanza <··@p-cos.net> writes:
> I actually had problems installing libsigsev on a OS X PowerPC system,
> so I always install clisp without it. (Don't remember the details,
> though...)
Yeah. For some reason, I always have a hard time convincing the
configure script that I do have libsigsegv. ISTR I managed to convince
it once but most of the times I give up and build it without libsigsegv.
--
Luís Oliveira
luismbo (@) gmail (.) com
http://student.dei.uc.pt/~lmoliv/
> * Luís Oliveira <·············@qrnqfcnz.pbz> [2006-07-18 01:07:16 +0100]:
>
> Pascal Costanza <··@p-cos.net> writes:
>> I actually had problems installing libsigsev on a OS X PowerPC system,
>> so I always install clisp without it. (Don't remember the details,
>> though...)
>
> Yeah. For some reason, I always have a hard time convincing the
> configure script that I do have libsigsegv. ISTR I managed to convince
> it once but most of the times I give up and build it without libsigsegv.
this is no good.
please try to investigate further and report your findings
http://clisp.cons.org/impnotes/faq.html#faq-bugs
thanks.
--
Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 5 (Bordeaux)
http://truepeace.org http://honestreporting.com http://mideasttruth.com
http://openvotingconsortium.org http://thereligionofpeace.com
will write code that writes code that writes code for food
Sam Steingold wrote:
>> * Luís Oliveira <·············@qrnqfcnz.pbz> [2006-07-18 01:07:16 +0100]:
>>
>> Pascal Costanza <··@p-cos.net> writes:
>>> I actually had problems installing libsigsev on a OS X PowerPC system,
>>> so I always install clisp without it. (Don't remember the details,
>>> though...)
>> Yeah. For some reason, I always have a hard time convincing the
>> configure script that I do have libsigsegv. ISTR I managed to convince
>> it once but most of the times I give up and build it without libsigsegv.
>
> this is no good.
For the sake of completeness, I have now been able to build clisp 2.39
successfully on Mac OS X 10.4.7, PowerPC G4 with the following incantation:
./configure --prefix=$HOME --build mac-osx-10.4.7
--with-libreadline-prefix=/sw --with-libiconv-prefix=/sw
--with-libintl-prefix=/sw --with-libgettext-prefix=/sw
I am using fink 0.8.1 for this. Installing libsigsegv from the original
sources is also necessary, and worked fine for me this time.
One reason why I couldn't initially build clisp was that
clisp-2.39.tar.gz was saved in a folder with the current date as a name
(in this case 19:07:2006, shown as 19/07/2006 in the Finder), due to the
Saft extension for Safari. The clisp build process doesn't seem to like
this.
There is still one failure in the test suite, which doesn't look too
problematic though.
Pascal
--
My website: http://p-cos.net
Closer to MOP & ContextL:
http://common-lisp.net/project/closer/
In article <··············@individual.net>,
Pascal Costanza <··@p-cos.net> wrote:
> Sam Steingold wrote:
> >> * Luís Oliveira <·············@qrnqfcnz.pbz> [2006-07-18 01:07:16 +0100]:
> >>
> >> Pascal Costanza <··@p-cos.net> writes:
> >>> I actually had problems installing libsigsev on a OS X PowerPC system,
> >>> so I always install clisp without it. (Don't remember the details,
> >>> though...)
> >> Yeah. For some reason, I always have a hard time convincing the
> >> configure script that I do have libsigsegv. ISTR I managed to convince
> >> it once but most of the times I give up and build it without libsigsegv.
> >
> > this is no good.
>
> For the sake of completeness, I have now been able to build clisp 2.39
> successfully on Mac OS X 10.4.7, PowerPC G4
Hm, didn't work for me under apparently identical conditions:
gcc -I/usr/local/include -g -O2 -W -Wswitch -Wcomment -Wpointer-arith
-Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare -O2
-DUNIX_BINARY_DISTRIB -DUNICODE -DDYNAMIC_FFI -DNO_GETTEXT -I. -I.. -c
calls.m.c -o calls.o
calls.c: In function `C_subr_posix_stream_options':
calls.c:256: warning: `value' might be used uninitialized in this
function
calls.c: In function `C_subr_posix_service':
calls.c:1326: warning: `se' might be used uninitialized in this function
calls.c: In function `C_subr_posix_stat_vfs':
calls.c:2111: error: storage size of `buf' isn't known
calls.c:2122: warning: implicit declaration of function `fstatvfs'
calls.c:2139: warning: implicit declaration of function `statvfs'
calls.c:2143: warning: left-hand operand of comma expression has no
effect
calls.c:2144: warning: left-hand operand of comma expression has no
effect
calls.c:2150: error: `fsblkcnt_t' undeclared (first use in this function)
calls.c:2150: error: (Each undeclared identifier is reported only once
calls.c:2150: error: for each function it appears in.)
calls.c:2150: warning: left-hand operand of comma expression has no
effect
calls.c:2151: warning: left-hand operand of comma expression has no
effect
calls.c:2152: warning: left-hand operand of comma expression has no
effect
calls.c:2160: error: `fsfilcnt_t' undeclared (first use in this function)
calls.c:2160: warning: left-hand operand of comma expression has no
effect
calls.c:2161: warning: left-hand operand of comma expression has no
effect
calls.c:2162: warning: left-hand operand of comma expression has no
effect
calls.c:2172: warning: left-hand operand of comma expression has no
effect
calls.c:2175: warning: left-hand operand of comma expression has no
effect
make[1]: *** [calls.o] Error 1
make: *** [syscalls] Error 2
rg
In article <·······························@news.gha.chartermi.net>,
Ron Garret <·········@flownet.com> wrote:
> In article <··············@individual.net>,
> Pascal Costanza <··@p-cos.net> wrote:
>
> > Sam Steingold wrote:
> > >> * Luís Oliveira <·············@qrnqfcnz.pbz> [2006-07-18 01:07:16
> > >> +0100]:
> > >>
> > >> Pascal Costanza <··@p-cos.net> writes:
> > >>> I actually had problems installing libsigsev on a OS X PowerPC system,
> > >>> so I always install clisp without it. (Don't remember the details,
> > >>> though...)
> > >> Yeah. For some reason, I always have a hard time convincing the
> > >> configure script that I do have libsigsegv. ISTR I managed to convince
> > >> it once but most of the times I give up and build it without libsigsegv.
> > >
> > > this is no good.
> >
> > For the sake of completeness, I have now been able to build clisp 2.39
> > successfully on Mac OS X 10.4.7, PowerPC G4
>
> Hm, didn't work for me under apparently identical conditions:
Correction: there is one difference in my setup: I'm not using Fink.
That could explain it. Still seems like a CLisp config bug that ought
to be fixed.
rg
Ron Garret wrote:
> In article <·······························@news.gha.chartermi.net>,
> Ron Garret <·········@flownet.com> wrote:
>
>> In article <··············@individual.net>,
>> Pascal Costanza <··@p-cos.net> wrote:
>>
>>> Sam Steingold wrote:
>>>>> * Luís Oliveira <·············@qrnqfcnz.pbz> [2006-07-18 01:07:16
>>>>> +0100]:
>>>>>
>>>>> Pascal Costanza <··@p-cos.net> writes:
>>>>>> I actually had problems installing libsigsev on a OS X PowerPC system,
>>>>>> so I always install clisp without it. (Don't remember the details,
>>>>>> though...)
>>>>> Yeah. For some reason, I always have a hard time convincing the
>>>>> configure script that I do have libsigsegv. ISTR I managed to convince
>>>>> it once but most of the times I give up and build it without libsigsegv.
>>>> this is no good.
>>> For the sake of completeness, I have now been able to build clisp 2.39
>>> successfully on Mac OS X 10.4.7, PowerPC G4
>> Hm, didn't work for me under apparently identical conditions:
>
> Correction: there is one difference in my setup: I'm not using Fink.
> That could explain it. Still seems like a CLisp config bug that ought
> to be fixed.
The implementations of some the required libraries (readline, gettext,
etc.) that Apple provides are incomplete / non-standard / whatever.
Clisp relies on standard GNU implementations. This is what fink
provides, so it's important that you use those versions. (DarwinPorts
apparently also does the job, as probably would manual installation of
the required libraries.)
It's probably also a good idea to use the newest XCode.
Pascal
--
My website: http://p-cos.net
Closer to MOP & ContextL:
http://common-lisp.net/project/closer/
On Wed, 19 Jul 2006 22:36:35 +0200, Pascal Costanza wrote:
> The implementations of some the required libraries (readline, gettext,
> etc.) that Apple provides are incomplete / non-standard / whatever.
> Clisp relies on standard GNU implementations. This is what fink
> provides, so it's important that you use those versions. (DarwinPorts
> apparently also does the job, as probably would manual installation of
> the required libraries.)
>
> It's probably also a good idea to use the newest XCode.
Here is a fink build-script for libsigsegv: libsigsegv.info
8<----------cut here----------------->
Package: libsigsegv
Architecture: powerpc
Version: 2.4
Revision: 1
License: GPL
Description: A library for handling page faults in user mode.
Maintainer: Guy Smiley <···@subgenuis.org>
Depends:
BuildDepends:
Source: http://ftp.gnu.org/gnu/libsigsegv/libsigsegv-2.4.tar.gz
Source-MD5: df0fe5f87a01e0b4b53b2cdfa8a730e1
GCC: 4.0
NoSetMAKEFLAGS: true
SetMAKEFLAGS: -j1
ConfigureParams:
Compilescript: <<
./configure %c
make
make check
<<
InstallScript: cd src; make DESTDIR=%d install
DescDetail: <<
A library for handling page faults in user mode.
<<
Homepage: http://libsigsegv.sourceforge.net/
8<----------cut here----------------->
To use it, place it in you local package tree.
Tested under 10.4.7, xcode-tools 23.
Usual disclaimers of damage to life and limb apply.
Matt
--
"You do not really understand something unless you can
explain it to your grandmother." — Albert Einstein.
> * Ron Garret <·········@sybjarg.pbz> [2006-07-19 13:26:47 -0700]:
>
> Hm, didn't work for me under apparently identical conditions:
>
> gcc -I/usr/local/include -g -O2 -W -Wswitch -Wcomment -Wpointer-arith
> -Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare -O2
> -DUNIX_BINARY_DISTRIB -DUNICODE -DDYNAMIC_FFI -DNO_GETTEXT -I. -I.. -c
> calls.m.c -o calls.o
> calls.c: In function `C_subr_posix_stream_options':
> calls.c:256: warning: `value' might be used uninitialized in this
> function
> calls.c: In function `C_subr_posix_service':
> calls.c:1326: warning: `se' might be used uninitialized in this function
> calls.c: In function `C_subr_posix_stat_vfs':
> calls.c:2111: error: storage size of `buf' isn't known
> calls.c:2122: warning: implicit declaration of function `fstatvfs'
> calls.c:2139: warning: implicit declaration of function `statvfs'
> calls.c:2143: warning: left-hand operand of comma expression has no
> effect
> calls.c:2144: warning: left-hand operand of comma expression has no
> effect
> calls.c:2150: error: `fsblkcnt_t' undeclared (first use in this function)
> calls.c:2150: error: (Each undeclared identifier is reported only once
> calls.c:2150: error: for each function it appears in.)
> calls.c:2150: warning: left-hand operand of comma expression has no
> effect
> calls.c:2151: warning: left-hand operand of comma expression has no
> effect
> calls.c:2152: warning: left-hand operand of comma expression has no
> effect
> calls.c:2160: error: `fsfilcnt_t' undeclared (first use in this function)
> calls.c:2160: warning: left-hand operand of comma expression has no
> effect
> calls.c:2161: warning: left-hand operand of comma expression has no
> effect
> calls.c:2162: warning: left-hand operand of comma expression has no
> effect
> calls.c:2172: warning: left-hand operand of comma expression has no
> effect
> calls.c:2175: warning: left-hand operand of comma expression has no
> effect
> make[1]: *** [calls.o] Error 1
> make: *** [syscalls] Error 2
the code that causes the errors is guarded by "ifdef HAVE_STATVFS".
http://www.opengroup.org/onlinepubs/009695399/functions/fstatvfs.html
where are these functions and structs declared on osx?
not in sys/statvfs.h I take it?
--
Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 5 (Bordeaux)
http://openvotingconsortium.org http://thereligionofpeace.com
http://mideasttruth.com http://palestinefacts.org http://ffii.org
There are no answers, only cross references.
In article <··············@quant8.janestcapital.quant>,
Sam Steingold <···@podval.org> wrote:
> the code that causes the errors is guarded by "ifdef HAVE_STATVFS".
> http://www.opengroup.org/onlinepubs/009695399/functions/fstatvfs.html
> where are these functions and structs declared on osx?
> not in sys/statvfs.h I take it?
Well, the situation is muddled now.
On Pascal's suggestion, I installed the latest XTools (the version on
the OSX Tiger Install DVD). That made most of the errors I was seeing
go away, but I was still getting this:
calls.c: In function 'C_subr_posix_stat_vfs':
calls.c:2111: error: storage size of 'buf' isn't known
So I did a make distclean (figuring that there was now a mismatch
between my config setup and my actual configuration) and started from
scratch, and now I'm back to an older problem, to wit, that it
absolutely refuses to recognize the fact that I do in fact have
libsigsegv installed (it's in /usr/local/lib), even when I specify
--with-libsigsegv-prefix. This problem has been reported by others as
well.
I still have to try rebuilding libsigsegv. It could be screwed up now
as a result of upgrading XTools. Stand by for a report.
rg
In article <·······························@news.gha.chartermi.net>,
Ron Garret <·········@flownet.com> wrote:
> I still have to try rebuilding libsigsegv. It could be screwed up now
> as a result of upgrading XTools. Stand by for a report.
I am pleased to report the successful compilation of CLisp 2.39 on a
fully stock OSX 10.4.7 (i.e. without using fink). The key steps (for
me) were an upgrade of XCode Tools and a subsequent clean rebuild of
libsigsegv.
One puzzle remains: I got a long list of warnings at the end of the form:
WARNING: FFI::LOOKUP-FOREIGN-FUNCTION: foreign function "rl_set_prompt"
does
not exist
Skip foreign function creation
which seems to indicate a problem linking to libreadline, and yet the
resulting CLisp does have readline. Go figure.
rg
> * Ron Garret <·········@sybjarg.pbz> [2006-07-20 10:33:07 -0700]:
>
> One puzzle remains: I got a long list of warnings at the end of the form:
>
> WARNING: FFI::LOOKUP-FOREIGN-FUNCTION: foreign function "rl_set_prompt"
> does
> not exist
> Skip foreign function creation
>
> which seems to indicate a problem linking to libreadline, and yet the
> resulting CLisp does have readline. Go figure.
these messages stem from building the readline module (which offers the
programmer a much finer control over the command line editing than the
default behavior) and they mean that you do not have the "latest and
greatest" readline.
--
Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 5 (Bordeaux)
http://ffii.org http://mideasttruth.com http://iris.org.il
http://truepeace.org http://camera.org http://thereligionofpeace.com
Parachute for sale, used once, never opened, small stain.
Pascal Costanza wrote:
> For the sake of completeness, I have now been able to build clisp 2.39
> successfully on Mac OS X 10.4.7, PowerPC G4 with the following incantation:
>
> ./configure --prefix=$HOME --build mac-osx-10.4.7
> --with-libreadline-prefix=/sw --with-libiconv-prefix=/sw
> --with-libintl-prefix=/sw --with-libgettext-prefix=/sw
>
> There is still one failure in the test suite, which doesn't look too
> problematic though.
Does FFI work? I've had problems building FFI on both PPC and Intel.
-kevin
Kevin Y. Kim wrote:
> Pascal Costanza wrote:
>> For the sake of completeness, I have now been able to build clisp 2.39
>> successfully on Mac OS X 10.4.7, PowerPC G4 with the following incantation:
>>
>> ./configure --prefix=$HOME --build mac-osx-10.4.7
>> --with-libreadline-prefix=/sw --with-libiconv-prefix=/sw
>> --with-libintl-prefix=/sw --with-libgettext-prefix=/sw
>>
>> There is still one failure in the test suite, which doesn't look too
>> problematic though.
>
> Does FFI work? I've had problems building FFI on both PPC and Intel.
If this is something that you have to ask for explicitly, then I simply
haven't tried. I have followed the instructions in the INSTALL
descriptions, without any changes to settings apart from the command
line given above.
The XCode version I use is 2.3, which you can download from
http://developer.apple.com
I am surprised by Ron's report that he didn't need fink (or something
similar) to build his system. This didn't work for me with past releases
of clisp.
Pascal
--
My website: http://p-cos.net
Closer to MOP & ContextL:
http://common-lisp.net/project/closer/
In article <··············@individual.net>,
Pascal Costanza <··@p-cos.net> wrote:
> Kevin Y. Kim wrote:
> > Pascal Costanza wrote:
> >> For the sake of completeness, I have now been able to build clisp 2.39
> >> successfully on Mac OS X 10.4.7, PowerPC G4 with the following incantation:
> >>
> >> ./configure --prefix=$HOME --build mac-osx-10.4.7
> >> --with-libreadline-prefix=/sw --with-libiconv-prefix=/sw
> >> --with-libintl-prefix=/sw --with-libgettext-prefix=/sw
> >>
> >> There is still one failure in the test suite, which doesn't look too
> >> problematic though.
> >
> > Does FFI work? I've had problems building FFI on both PPC and Intel.
>
> If this is something that you have to ask for explicitly, then I simply
> haven't tried. I have followed the instructions in the INSTALL
> descriptions, without any changes to settings apart from the command
> line given above.
>
> The XCode version I use is 2.3, which you can download from
> http://developer.apple.com
>
> I am surprised by Ron's report that he didn't need fink (or something
> similar) to build his system. This didn't work for me with past releases
> of clisp.
All of my CLisp builds have been finkless and ultimately (sometimes
after a bit of head-scratching) successful. My previous build was 2.38
under OSX 10.3, which went smoothly except for trying to convince CLisp
that I really did have libsigsegv installed. (I think I ultimately
punted and decided to just live without generational GC since I mostly
use CLisp for compatibility tests rather than real work due to its lack
of threads.)
I vaguely remember some actual problems building under OSX in the dim
and distant past, but they were quickly fixed by the CLisp developers.
rg
In article <·······························@news.gha.chartermi.net>,
Ron Garret <·········@flownet.com> wrote:
> I mostly use CLisp for compatibility tests
I note with a certain wry amusement that (ignore-errors (function
ignore-errors)) returns NIL in CLisp even in compiled code.
rg
Ron Garret <·········@flownet.com> wrote:
+---------------
| Ron Garret <·········@flownet.com> wrote:
| > I mostly use CLisp for compatibility tests
|
| I note with a certain wry amusement that (ignore-errors (function
| ignore-errors)) returns NIL in CLisp even in compiled code.
+---------------
You need to look at the *second* value, too, which per the CLHS
entry for IGNORE-ERRORS contains the condition resulting from the
error [if any]. E.g., in CLISP:
[1]> (ignore-errors (function ignore-errors))
NIL ;
#<SYSTEM::SIMPLE-UNDEFINED-FUNCTION #x203FB779>
[2]>
or CMUCL:
cmu> (ignore-errors (function ignore-errors))
; In: IGNORE-ERRORS #'IGNORE-ERRORS
; #'IGNORE-ERRORS
; Error: Found macro name IGNORE-ERRORS as the argument to FUNCTION.
;
NIL
#<KERNEL:SIMPLE-PROGRAM-ERROR {589054B5}>
cmu>
Though compiled CLISP's behavior is a bit "different":
[3]> (defun foo () (ignore-errors (function ignore-errors)))
FOO
[4]> (compile *)
FOO ;
NIL ;
NIL
[5]> (foo)
#<MACRO #<COMPILED-CLOSURE IGNORE-ERRORS>>
[6]>
CMUCL, while complaining mightily during the COMPILE, still returns
two reasonable values in the compiled case:
cmu> (defun foo () (ignore-errors (function ignore-errors)))
FOO
cmu> (compile *)
; In: LAMBDA NIL
; #'IGNORE-ERRORS
; Error: Found macro name IGNORE-ERRORS as the argument to FUNCTION.
;
; Compiling LAMBDA NIL:
; Compiling Top-Level Form:
; Compilation unit finished.
; 1 error
FOO
T
T
cmu> (foo)
NIL
#<KERNEL:SIMPLE-PROGRAM-ERROR {58947CA5}>
cmu>
-Rob
-----
Rob Warnock <····@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
On 2006-07-20 14:31:59 -0400, Ron Garret <·········@flownet.com> said:
> All of my CLisp builds have been finkless and ultimately (sometimes
> after a bit of head-scratching) successful.
I always wuss out and just wait for darwinports to be updated.
Novus
Pascal Costanza wrote:
> bradb wrote:
> > Javier wrote:
> >> bradb wrote:
> >>> Javier wrote:
> >>>> Thank you.
> >>>>
> >>>> When is CLISP going to support Intel Mac OSX?
> >>> I happened to compile CLISP from CVS yesteday on Intel OSX without any
> >>> problems. I couldn't get libsigsev to work right away, so I forced
> >>> CLISP to ignore it.
> >>> CLISP appeared to install and work correctly, I could run the Slime
> >>> backend and evaluate forms.
> >> Sure, but without libsigsev there is not generation GC (that is said in
> >> the INSTALL file).
> > Does that mean there is no alternative GC? I noticed the warnings upon
> > install about no generational GC, but I would expect if there was no GC
> > at all I would get big screaming letters telling me that CLISP was
> > cripped by lack of a GC.
> > libsigsegv appears to have worked on Darwin i686 as recently as 2.4, I
> > can't imagine that there are huge changes/bugs preventing libsegsegv
> > from working on OSX Intel.
>
> I actually had problems installing libsigsev on a OS X PowerPC system,
> so I always install clisp without it. (Don't remember the details,
> though...)
Seems to have built OK with DarwinPorts for me, on PPC
Sometime I tweak the portfile to try different build options
BUT libiconv is back at 1.9.2_1 (from 1.10_1, never been able to
upgrade)
·····@yahoo.com wrote:
> Pascal Costanza wrote:
>> bradb wrote:
>>> Javier wrote:
>>>> bradb wrote:
>>>>> Javier wrote:
>>>>>> Thank you.
>>>>>>
>>>>>> When is CLISP going to support Intel Mac OSX?
>>>>> I happened to compile CLISP from CVS yesteday on Intel OSX without any
>>>>> problems. I couldn't get libsigsev to work right away, so I forced
>>>>> CLISP to ignore it.
>>>>> CLISP appeared to install and work correctly, I could run the Slime
>>>>> backend and evaluate forms.
>>>> Sure, but without libsigsev there is not generation GC (that is said in
>>>> the INSTALL file).
>>> Does that mean there is no alternative GC? I noticed the warnings upon
>>> install about no generational GC, but I would expect if there was no GC
>>> at all I would get big screaming letters telling me that CLISP was
>>> cripped by lack of a GC.
>>> libsigsegv appears to have worked on Darwin i686 as recently as 2.4, I
>>> can't imagine that there are huge changes/bugs preventing libsegsegv
>>> from working on OSX Intel.
>> I actually had problems installing libsigsev on a OS X PowerPC system,
>> so I always install clisp without it. (Don't remember the details,
>> though...)
>
> Seems to have built OK with DarwinPorts for me, on PPC
> Sometime I tweak the portfile to try different build options
>
> BUT libiconv is back at 1.9.2_1 (from 1.10_1, never been able to
> upgrade)
DarwinPorts doesn't seem to have 2.39 yet. Do you use the official 2.39
release of clisp, or the cvs version? I am asking this because the 2.39
release doesn't build on my machine, but the cvs version does (which I
find quite strange).
Pascal
--
My website: http://p-cos.net
Closer to MOP & ContextL:
http://common-lisp.net/project/closer/
Sam Steingold wrote:
> 2.39 (2006-07-16)
You know, you guys really should have some kind of pre-release staging
for CLISP. It seems that time a release is put out, people download it
and find silly little problems which are then fixed in CVS.
The result is that the version history of CLISP is littered with
official releases that are subtly broken.
Put out a 2.38.95 or 2.38-pre-1 or something. Then if nothing goes
wrong, you just tweak the version numbers and make it into 2.39.
Otherwise, you can make a pre-2 and so on.
Not everyone sits there going "cvs up" everyday just in case you might
release something.
Or, heck, if candidate tarballs are too much trouble, how about just
doing the branching in CVS and let people know about it:
``We have made a release branch called REL_2_39. Try it, give us
feedback.''
People who care about that release will ``cvs up'' from the branch and
do some testing, having an opportunity to provide feedback. Fixes are
done in the branch, and downmerged to the trunk where new work can
already be going on.
It's the early warning that is more important than how the details are
managed. It can be done without branching by imposing a bugfix-only
phase on the trunk where the candidates are tagged, or whatever.
> * Kaz Kylheku <········@tznvy.pbz> [2006-07-17 11:38:21 -0700]:
>
> Sam Steingold wrote:
>> 2.39 (2006-07-16)
>
> You know, you guys really should have some kind of pre-release staging
> for CLISP. It seems that time a release is put out, people download it
> and find silly little problems which are then fixed in CVS.
what silly problems with 2.39 have you found?
> ``We have made a release branch called REL_2_39. Try it, give us
> feedback.''
last week I wrote to clisp-list that 2.39 was imminent and asked for
volunteers to build woe32 packages.
--
Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 5 (Bordeaux)
http://pmw.org.il http://camera.org http://iris.org.il
http://mideasttruth.com http://memri.org http://openvotingconsortium.org
WinWord 6.0 UNinstall: Not enough disk space to uninstall WinWord
Sam Steingold schrieb:
> ANSI Common Lisp is a high-level, general-purpose programming language.
> GNU CLISP is a Common Lisp implementation by Bruno Haible of Karlsruhe
> University and Michael Stoll of Munich University, both in Germany.
> It mostly supports the Lisp described in the ANSI Common Lisp standard.
> It runs on most GNU and Unix systems (GNU/Linux, FreeBSD, NetBSD, OpenBSD,
> Solaris, Tru64, HP-UX, BeOS, NeXTstep, IRIX, AIX and others) and on
> other systems (Windows NT/2000/XP, Windows 95/98/ME) and needs only
> 4 MB of RAM.
> It is Free Software and may be distributed under the terms of GNU GPL,
> while it is possible to distribute commercial proprietary applications
> compiled with GNU CLISP.
> The user interface comes in English, German, French, Spanish, Dutch,
> Russian and Danish, and can be changed at run time.
> GNU CLISP includes an interpreter, a compiler, a debugger, CLOS, MOP,
> a foreign language interface, sockets, i18n, fast bignums and more.
> An X11 interface is available through CLX, Garnet, CLUE/CLIO.
> GNU CLISP runs Maxima, ACL2 and many other Common Lisp packages.
>
> More information at
> <http://clisp.cons.org/>,
> <http://www.clisp.org/>,
> <http://www.gnu.org/software/clisp/> and
> <http://clisp.sourceforge.net/>.
> Sources and selected binaries are available by anonymous ftp from
> <ftp://ftp.gnu.org/pub/gnu/clisp/>
> and its mirrors.
>
> 2.39 (2006-07-16)
> =================
>
> User visible changes
> --------------------
>
> * SAVEINITMEM now accepts :SCRIPT argument that disables interpreting
> the first positional argument as the script name; and :DOCUMENTATION
> argument that is printed by the new -help-image command line option.
> See <http://clisp.cons.org/impnotes/image.html> and
> <http://clisp.cons.org/impnotes/clisp.html#opt-help-image> for details.
>
> * FFI:UINT64 and FFI:SINT64 are now compatible with C's long long type.
>
> * Stack overflow detection and recovery finally work fine on Unix.
> Libsigsegv is required for this, on all platforms (including MS-Windows).
> CLISP should neither exit nor crash under infinite recursion.
> If your distribution has CLISP compiled without libsigsegv, report
> the missing feature to its maintainer.
> Note that libsigsegv 2.4 is required, there are bugs in libsigsegv 2.3!
>
> * It is now possible to specify the default method-combination of a generic
> function, to be used when the DEFGENERIC form does not specify the
> :METHOD-COMBINATION explicitly, through a default initarg specification
> for the :METHOD-COMBINATION keyword on the generic function class.
>
> * Readline completion works with non 1:1 terminal encodings, e.g. UTF-8.
>
> * WITH-KEYBOARD works with a Unix tty even when SLIME hijacks *TERMINAL-IO*.
>
> * I/O operations on Win32 are now much faster.
>
> * New functions: POSIX:FFS, POSIX:PATHCONF.
>
> * Infrastructure:
> + Top-level configure now accepts a new option --with-gmalloc to use the
> GNU malloc implementation instead of the one supplied by libc.
> You may need it on older HP-UX and newer OpenBSD systems.
> See file unix/PLATFORMS for more information.
> + The value of the environment variable CFLAGS is respected by configure.
>
> * Bug fixes:
> + SOCKET:SOCKET-SERVER :INTERFACE now behaves as documented.
> + EXT:READ-BYTE-NO-HANG and SOCKET:SOCKET-STATUS used to hang on
> buffered binary sockets.
> + Allow DESTRUCTURING-BIND (a . b) with circular and dotted lists.
> + ADJUST-ARRAY of zero length adjustable string now works.
> + TIME now reports correct results when the heap grows over 4GB.
> + RAWSOCK functions now handle :START/:END arguments correctly.
> + BDB:DBC-GET now accepts :READ-COMMITTED and :READ-UNCOMMITTED.
> + POSIX:GROUP-INFO and POSIX:USER-INFO now handle errors correctly.
>
> * Portability:
> + Support DragonFly BSD.
>
> --
> Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 5 (Bordeaux)
> http://memri.org http://mideasttruth.com http://camera.org
> http://iris.org.il http://palestinefacts.org http://jihadwatch.org
> Please wait, MS Windows are preparing the blue screen of death.
Sam Steingold schrieb:
> ANSI Common Lisp is a high-level, general-purpose programming language.
> GNU CLISP is a Common Lisp implementation by Bruno Haible of Karlsruhe
> University and Michael Stoll of Munich University, both in Germany.
> It mostly supports the Lisp described in the ANSI Common Lisp standard.
> It runs on most GNU and Unix systems (GNU/Linux, FreeBSD, NetBSD, OpenBSD,
> Solaris, Tru64, HP-UX, BeOS, NeXTstep, IRIX, AIX and others) and on
> other systems (Windows NT/2000/XP, Windows 95/98/ME) and needs only
> 4 MB of RAM.
> It is Free Software and may be distributed under the terms of GNU GPL,
> while it is possible to distribute commercial proprietary applications
> compiled with GNU CLISP.
> The user interface comes in English, German, French, Spanish, Dutch,
> Russian and Danish, and can be changed at run time.
> GNU CLISP includes an interpreter, a compiler, a debugger, CLOS, MOP,
> a foreign language interface, sockets, i18n, fast bignums and more.
> An X11 interface is available through CLX, Garnet, CLUE/CLIO.
> GNU CLISP runs Maxima, ACL2 and many other Common Lisp packages.
>
> More information at
> <http://clisp.cons.org/>,
> <http://www.clisp.org/>,
> <http://www.gnu.org/software/clisp/> and
> <http://clisp.sourceforge.net/>.
> Sources and selected binaries are available by anonymous ftp from
> <ftp://ftp.gnu.org/pub/gnu/clisp/>
> and its mirrors.
>
> 2.39 (2006-07-16)
> =================
>
> User visible changes
> --------------------
>
> * SAVEINITMEM now accepts :SCRIPT argument that disables interpreting
> the first positional argument as the script name; and :DOCUMENTATION
> argument that is printed by the new -help-image command line option.
> See <http://clisp.cons.org/impnotes/image.html> and
> <http://clisp.cons.org/impnotes/clisp.html#opt-help-image> for details.
>
> * FFI:UINT64 and FFI:SINT64 are now compatible with C's long long type.
>
> * Stack overflow detection and recovery finally work fine on Unix.
> Libsigsegv is required for this, on all platforms (including MS-Windows).
> CLISP should neither exit nor crash under infinite recursion.
> If your distribution has CLISP compiled without libsigsegv, report
> the missing feature to its maintainer.
> Note that libsigsegv 2.4 is required, there are bugs in libsigsegv 2.3!
>
> * It is now possible to specify the default method-combination of a generic
> function, to be used when the DEFGENERIC form does not specify the
> :METHOD-COMBINATION explicitly, through a default initarg specification
> for the :METHOD-COMBINATION keyword on the generic function class.
>
> * Readline completion works with non 1:1 terminal encodings, e.g. UTF-8.
>
> * WITH-KEYBOARD works with a Unix tty even when SLIME hijacks *TERMINAL-IO*.
>
> * I/O operations on Win32 are now much faster.
>
> * New functions: POSIX:FFS, POSIX:PATHCONF.
>
> * Infrastructure:
> + Top-level configure now accepts a new option --with-gmalloc to use the
> GNU malloc implementation instead of the one supplied by libc.
> You may need it on older HP-UX and newer OpenBSD systems.
> See file unix/PLATFORMS for more information.
> + The value of the environment variable CFLAGS is respected by configure.
>
> * Bug fixes:
> + SOCKET:SOCKET-SERVER :INTERFACE now behaves as documented.
> + EXT:READ-BYTE-NO-HANG and SOCKET:SOCKET-STATUS used to hang on
> buffered binary sockets.
> + Allow DESTRUCTURING-BIND (a . b) with circular and dotted lists.
> + ADJUST-ARRAY of zero length adjustable string now works.
> + TIME now reports correct results when the heap grows over 4GB.
> + RAWSOCK functions now handle :START/:END arguments correctly.
> + BDB:DBC-GET now accepts :READ-COMMITTED and :READ-UNCOMMITTED.
> + POSIX:GROUP-INFO and POSIX:USER-INFO now handle errors correctly.
>
> * Portability:
> + Support DragonFly BSD.
>
> --
> Sam Steingold (http://www.podval.org/~sds) on Fedora Core release 5 (Bordeaux)
> http://memri.org http://mideasttruth.com http://camera.org
> http://iris.org.il http://palestinefacts.org http://jihadwatch.org
> Please wait, MS Windows are preparing the blue screen of death.
Thank you for the great job.
keyboard
On Mon, 17 Jul 2006 13:28:40 -0400, Sam Steingold <···@podval.org> tried
to confuse everyone with this message:
>
>* I/O operations on Win32 are now much faster.
>
Ooh, I can feel it! My Ltk app runs much smoother (so I don't have to
optimise that code :)). Thank you very much.
--
|Don't believe this - you're not worthless ,gr---------.ru
|It's us against millions and we can't take them all... | ue il |
|But we can take them on! | @ma |
| (A Wilhelm Scream - The Rip) |______________|