CLISP - an ANSI Common Lisp

Sam wrote:
> for now, adding
>
> #if defined(__MINGW32__)
> #define ENABLE_NLS 1
> #endif
>
> to win32.d should fix the immediate problem.

It would be better if ENABLE_NLS got defined through the CFLAGS
in the Makefile, rather than through win32.d. Otherwise gettext would
become a build requirement on Woe32. It's better to not make this a
requirement but leave it optional.

Bruno

> * Bruno Haible <oehab@...> [2005-02-25 16:40:16 +0100]:
>
> Sam wrote:
>> for now, adding
>>
>> #if defined(__MINGW32__)
>> #define ENABLE_NLS 1
>> #endif
>>
>> to win32.d should fix the immediate problem.
>
> It would be better if ENABLE_NLS got defined through the CFLAGS in the
> Makefile, rather than through win32.d. Otherwise gettext would become
> a build requirement on Woe32. It's better to not make this a
> requirement but leave it optional.

better yet, use autoconf - include unixconf.h.

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
 <http://www.camera.org> <http://www.mideasttruth.com/> <http://www.memri.org/>
<http://www.honestreporting.com> <http://pmw.org.il/> <http://www.iris.org.il>
MS Windows: error: the operation completed successfully.

Sam Steingold:

> > build-full/clisp.h
> > ...
> > #define ENABLE_NLS 1
> > ...
> > #define GETTEXT(english) english
> > ...
> >
> > build-full/unixconf.h: (this for win32?)
>
> alas, unixconf.h, although generated on all platforms, is only used on
> unix (for the sake of msvc users).
> it really should change, but this is not trivial.
> if you want to work on this, change lispbibl.d around
> "# Determine properties of compiler and environment:"
> so that __MINGW32__ is the same as UNIX and see what happens
> (you will probably have to put much of win32.d under "#ifndef 
> __MINGW32__")
>
>
>
> for now, adding
>
> #if defined(__MINGW32__)
> #define ENABLE_NLS 1
> #endif
>
> to win32.d should fix the immediate problem.
>
Has added.
Problem is not fixed:
in clisp.h line has remained:
#define GETTEXT(english) english

clisp from cvs-tarball (not from cvs head)

-- 
WBR, Yaroslav Kavenchuk.

> * Yaroslav Kavenchuk <xnirapuhx@...> [2005-02-28 17:41:12 +0200]:
>
> Sam Steingold:
>
>> > build-full/clisp.h
>> > ...
>> > #define ENABLE_NLS 1
>> > ...
>> > #define GETTEXT(english) english
>> > ...
>> >
>> > build-full/unixconf.h: (this for win32?)
>>
>> alas, unixconf.h, although generated on all platforms, is only used on
>> unix (for the sake of msvc users).
>> it really should change, but this is not trivial.
>> if you want to work on this, change lispbibl.d around
>> "# Determine properties of compiler and environment:"
>> so that __MINGW32__ is the same as UNIX and see what happens
>> (you will probably have to put much of win32.d under "#ifndef
>> __MINGW32__")
>>
>>
>>
>> for now, adding
>>
>> #if defined(__MINGW32__)
>> #define ENABLE_NLS 1
>> #endif
>>
>> to win32.d should fix the immediate problem.
>>
> Has added.
> Problem is not fixed:
> in clisp.h line has remained:
> #define GETTEXT(english) english

well, I guess the trivial tweak did not work.

how about the appended patch on top of CVS head?

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.iris.org.il> <http://www.camera.org> <http://www.mideasttruth.com/>
<http://www.memri.org/> <http://www.honestreporting.com> <http://pmw.org.il/>
Daddy, why doesn't this magnet pick up this floppy disk?


Index: src/win32.d
===================================================================
RCS file: /cvsroot/clisp/clisp/src/win32.d,v
retrieving revision 1.51
diff -u -w -r1.51 win32.d
--- src/win32.d	19 Jan 2005 14:47:59 -0000	1.51
+++ src/win32.d	28 Feb 2005 16:51:43 -0000
@@ -20,7 +20,7 @@
 #endif
 
 /* for _clisp.c */
-#define STDC_HEADERS
+#define STDC_HEADERS 1
 #define HAVE_PERROR_DECL
 
 /* Declaration of operating system functions */
Index: src/lispbibl.d
===================================================================
RCS file: /cvsroot/clisp/clisp/src/lispbibl.d,v
retrieving revision 1.616
diff -u -w -r1.616 lispbibl.d
--- src/lispbibl.d	24 Feb 2005 19:11:54 -0000	1.616
+++ src/lispbibl.d	28 Feb 2005 16:51:44 -0000
@@ -312,11 +312,11 @@
 
 
 # Determine properties of compiler and environment:
-#if defined(UNIX)
+#if defined(UNIX) || defined(__MINGW32__)
   #include "unixconf.h"  # configuration generated by configure
   #include "intparam.h"  # integer-type characteristics created by the machine
   #include "floatparam.h" # floating-point type characteristics
-#elif defined(WIN32)
+#elif defined(WIN32) && !defined(__MINGW32__)
   #include "version.h"          /* defines PACKAGE_* */
   #define char_bitsize 8
   #define short_bitsize 16
@@ -348,6 +348,8 @@
   #define SHLIB_ADDRESS_RANGE 0
   #define STACK_ADDRESS_RANGE ~0UL
   #define ICONV_CONST
+#else
+  #error "where is the configuration for your platform?"
 #endif

Sam Steingold:

> >> for now, adding
> >>
> >> #if defined(__MINGW32__)
> >> #define ENABLE_NLS 1
> >> #endif
> >>
> >> to win32.d should fix the immediate problem.
> >>
> > Has added.
> > Problem is not fixed:
> > in clisp.h line has remained:
> > #define GETTEXT(english) english
>
> well, I guess the trivial tweak did not work.
>
> how about the appended patch on top of CVS head?
>
top of CVS head (without patch):

$ ./configure --with-mingw --build build-mingw
....
gcc -mno-cygwin -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit 
-Wreturn-type  -Wno-sign-compare -O2 -fexpensive-optimizations -D_WIN32 
-DUNICODE -DNO_TERMCAP_NCURSES -DDYNAMIC_FFI -DNO_GETTEXT -I. -x none 
spvw.o spvwtabf.o spvwtabs.o spvwtabo.o eval.o control.o encoding.o 
pathname.o stream.o socket.o io.o array.o hashtabl.o list.o package.o 
record.o weak.o sequence.o charstrg.o debug.o error.o misc.o time.o 
predtype.o symbol.o lisparit.o i18n.o foreign.o win32aux.o built.o 
ari80386.o modules.o libcharset.a libavcall.a libcallback.a  -luser32 
-lws2_32 -lole32 -loleaut32 -luuid -L/usr/local/lib -lsigsegv -o lisp.exe
rm -f init.lisp
ln ../src/init.lisp init.lisp
....
rm -f deprecated.lisp
ln ../src/deprecated.lisp deprecated.lisp
rm -f interpreted.mem
lisp.exe -B . -Efile UTF-8 -Eterminal UTF-8 -norc -m 1400KW -x "(and 
(load \"init.lisp\") (sys::%saveinitmem) (ext::exit)) (ext::exit t)"
make: *** [interpreted.mem] Error 5


What error - I do not know.
Popup message: "lisp.exe has caused a error and it will be closed. It is 
necessary to restart the program." (my translate from russian)

-- 
WBR, Yaroslav Kavenchuk.

> * Yaroslav Kavenchuk <xnirapuhx@...> [2005-02-28 19:17:14 +0200]:
>
> make: *** [interpreted.mem] Error 5

./configure --with-debug --with-mingw --build build-g
run the failing command under gdb.


-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.mideasttruth.com/> <http://www.memri.org/> <http://www.camera.org>
<http://www.honestreporting.com> <http://pmw.org.il/>
Growing Old is Inevitable; Growing Up is Optional.

Sam Steingold:

> > make: *** [interpreted.mem] Error 5
>
> ./configure --with-debug --with-mingw --build build-g
> run the failing command under gdb.
>
`$ ./configure --with-debug...' do not generate error.

-- 
WBR, Yaroslav Kavenchuk.

Sam Steingold:

> > make: *** [interpreted.mem] Error 5
>
> ./configure --with-debug --with-mingw --build build-g
> run the failing command under gdb.
>
from CVS head (01.03.2005) no error

-- 
WBR, Yaroslav Kavenchuk.

Sam Steingold:

> how about the appended patch on top of CVS head?
>
$ ./configure --with-mingw --build build-patch
....
mkdir data
cd data && ln ../../utils/unicode/UnicodeDataFull.txt .
cd data && ln ../../src/clhs.txt .
gcc -mno-cygwin -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit 
-Wreturn-type  -Wno-sign-compare -O2 -fexpensive-optimizations -D_WIN32 
-DUNICODE -DNO_TERMCAP_NCURSES -DDYNAMIC_FFI -DNO_GETTEXT -I. -x none 
spvw.o spvwtabf.o spvwtabs.o spvwtabo.o eval.o control.o encoding.o 
pathname.o stream.o socket.o io.o array.o hashtabl.o list.o package.o 
record.o weak.o sequence.o charstrg.o debug.o error.o misc.o time.o 
predtype.o symbol.o lisparit.o i18n.o foreign.o win32aux.o built.o 
ari80386.o modules.o libcharset.a libavcall.a libcallback.a  -luser32 
-lws2_32 -lole32 -loleaut32 -luuid -L/usr/local/lib -lsigsegv -o lisp.exe
stream.o(.text+0x3e43):stream.c: undefined reference to 
`_imp__libiconv_open'
stream.o(.text+0x3e5c):stream.c: undefined reference to `errno'
stream.o(.text+0x3ecf):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x3ef2):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x4151):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x415a):stream.c: undefined reference to `errno'
stream.o(.text+0x41c3):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x421b):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x4225):stream.c: undefined reference to `errno'
stream.o(.text+0x423a):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x42cf):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x42d7):stream.c: undefined reference to `errno'
stream.o(.text+0x44d3):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4526):stream.c: undefined reference to `errno'
stream.o(.text+0x45b8):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x4601):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x460c):stream.c: undefined reference to `errno'
stream.o(.text+0x4624):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x4859):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4861):stream.c: undefined reference to `errno'
stream.o(.text+0x48fd):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4959):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4962):stream.c: undefined reference to `errno'
stream.o(.text+0x4979):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x49bb):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x4a1e):stream.c: undefined reference to `errno'
stream.o(.text+0x4ae9):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4af1):stream.c: undefined reference to `errno'
stream.o(.text+0x4b6c):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4d69):stream.c: undefined reference to `errno'
stream.o(.text+0x4df9):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4e01):stream.c: undefined reference to `errno'
stream.o(.text+0x4e7c):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4eb8):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x4ec1):stream.c: undefined reference to `errno'
stream.o(.text+0x4ed8):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x4ef5):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x4f5d):stream.c: undefined reference to `errno'
stream.o(.text+0x50da):stream.c: undefined reference to `errno'
stream.o(.text+0x513d):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x5214):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x52bb):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x554b):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x5570):stream.c: undefined reference to 
`_imp__libiconv_close'
stream.o(.text+0x6f3a):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x6f83):stream.c: undefined reference to `errno'
stream.o(.text+0x8b8a):stream.c: undefined reference to `_imp__libiconv'
stream.o(.text+0x8be2):stream.c: undefined reference to `errno'
collect2: ld returned 1 exit status
make: *** [lisp.exe] Error 1

-- 
WBR, Yaroslav Kavenchuk.

Yaroslav Kavenchuk:

> make: *** [lisp.exe] Error 1
>
I'm sorry.
I have forget to apply own patch for makemake.in (if it is necessary):

#########################################
$ diff -c -u makemake.in makemake.in.new
diff: conflicting specifications of output style
--- makemake.in Tue Mar  1 15:03:37 2005
+++ makemake.in.new     Tue Mar  1 15:02:45 2005
@@ -1532,7 +1532,17 @@
   fi
 fi
 if [ $TOS = win32 ] ; then
-  XCFLAGS=$XCFLAGS' -DNO_GETTEXT'
+  if [ "${with_gettext}" = no ]; then
+    XCFLAGS=$XCFLAGS' -DNO_GETTEXT'
+  else
+    if [ @USE_NLS@ = yes ] ; then
+      USE_GETTEXT=yes
+      LIBS='@LIBINTL@ '$LIBS
+      XCL_GETTEXTLIB=$XCL_GETTEXTLIB' locale'
+    else
+      XCFLAGS=$XCFLAGS' -DNO_GETTEXT'
+    fi
+  fi
 fi
 
 if test -z "$LIBSIGSEGV"; then
###############################################

but result same: forget "-liconv" and it is not known "errno" in `stream.c'.

-- 
WBR, Yaroslav Kavenchuk.

> * Yaroslav Kavenchuk <xnirapuhx@...> [2005-03-01 10:46:26 +0200]:
>
> stream.o(.text+0x3e5c):stream.c: undefined reference to `errno'

fixed in the CVS head.

> stream.o(.text+0x3ecf):stream.c: undefined reference to
> _imp__libiconv_close'

for some reason "-liconv" is not in LIBS.
why?
what is LIBICONV in build/makemake?
it should be "-liconv" or something like that.

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.iris.org.il> <http://www.honestreporting.com>
<http://www.camera.org> <http://www.mideasttruth.com/> <http://pmw.org.il/>
When you are arguing with an idiot, your opponent is doing the same.

> * Yaroslav Kavenchuk <xnirapuhx@...> [2005-03-01 17:38:26 +0200]:
>
> but result same: forget "-liconv" and it is not known "errno" in `stream.c'.

should be fixed in the CVS head.

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.honestreporting.com> <http://www.iris.org.il>
<http://pmw.org.il/> <http://www.mideasttruth.com/> <http://www.camera.org>
The difference between genius and stupidity is that genius has its limits.

Sam Steingold:

> > but result same: forget "-liconv" and it is not known "errno" in 
> `stream.c'.
>
> should be fixed in the CVS head.
>
Yes, but...

CVS head (02.03.05)
./configure --with-mingw --build build-mingw
....
gcc -mno-cygwin -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit 
-Wreturn-type  -Wno-sign-compare -O2 -fexpensive-optimizations -D_WIN32 
-DUNICODE -DNO_TERMCAP_NCURSES -DDYNAMIC_FFI -I. -x none spvw.o 
spvwtabf.o spvwtabs.o spvwtabo.o eval.o control.o encoding.o pathname.o 
stream.o socket.o io.o array.o hashtabl.o list.o package.o record.o 
weak.o sequence.o charstrg.o debug.o error.o misc.o time.o predtype.o 
symbol.o lisparit.o i18n.o foreign.o win32aux.o built.o ari80386.o 
modules.o -lintl libcharset.a libavcall.a libcallback.a  -luser32 
-lws2_32 -lole32 -loleaut32 -luuid -liconv -L/usr/local/lib -lsigsegv -o 
lisp.exe
stream.o(.text+0x7093):stream.c: undefined reference to `errno'
stream.o(.text+0x8d02):stream.c: undefined reference to `errno'
collect2: ld returned 1 exit status
make: *** [lisp.exe] Error 1

-- 
WBR, Yaroslav Kavenchuk.

From: "Hoehle, Joerg-Cyril" <Joerg-Cyril.Hoehle@...>
   Date: Tue, 1 Mar 2005 13:52:48 +0100 

   Hi,
   >Using SLIME, I was able to load and compile koza.lisp.  Then I saved
   >the result to a new image, koza.mem and changed clisp.bat to
   [...]
   >If I start SLIME instead, I get the endless
   >*** - Program Stack Overflow. RESET

   I have little experience with SLIME and no idea what causes this endless loop, but I remember some posting to the slime mailing list talking about problems with SWANK (the CL part of slime) in the server and restarting.

   Maybe that note is old, but who knows?
   Therefore, I'd recommend you try this:
   a) now that you know how to build koza with SLIME, go into a bare clisp (non slime), load the compiled koza from there and save the image.
   b) start SLIME with that image, loading swank anew for each session

   Maybe that could work?
   BTW, do you currently load swank in clisp in each session with the original clisp image or did you generate your own image with swank in it?


The koza image does include swank now that I think about it, since I
built it via SLIME.  That indeed could be the problem.  I'll try a
rebuild.

From: "Hoehle, Joerg-Cyril" <Joerg-Cyril.Hoehle@...>
   Date: Tue, 1 Mar 2005 13:52:48 +0100 

   >Using SLIME, I was able to load and compile koza.lisp.  Then I saved
   >the result to a new image, koza.mem and changed clisp.bat to
   [...]
   >If I start SLIME instead, I get the endless
   >*** - Program Stack Overflow. RESET

   ...talking about problems with SWANK (the CL part of slime) in the server and restarting.

   Therefore, I'd recommend you try this:
   a) ...go into a bare clisp (non slime), load the compiled koza from there and save the image.


That was indeed the problem.  SWANK cannot be re-loaded and must be
kept out of the saved image (or you need to customize your slime
startup).  Thanks.

> * GP lisper <scu@...> [2005-02-28 18:14:34 -0800]:
>
> The editbin program might help, if I knew where to get it.

google for it.
you _can_ get it (or an equivalent).

> Using SLIME, I was able to load and compile koza.lisp.

you probably mean "compile and load".

> with a load and compile of koza everytime I want to use it

you have to compile it just once, and then load the compiled (*.fas)
file(s)

I don't use swank, slime, koza &c, so I cannot say anything else.

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.memri.org/> <http://www.camera.org>
<http://www.mideasttruth.com/> <http://pmw.org.il/> <http://www.iris.org.il>
The program isn't debugged until the last user is dead.

Sam wrote:
> > [2]> (language-information)
> >
> > *** - invalid byte sequence #xCF #xED in CHARSET:UTF-8 conversion
> > The following restarts are available:
> > ABORT          :R1      ABORT
> > Break 1 [3]>
> >
> > clisp option "-E xxx" where xxx is CP1251, CP866 or UTF-8 does not
> > influence result

The macro res_to_obj() in modules/i18n/gettext.c uses UTF-8 always. I think
it should better use *MISC-ENCODING* instead.

Bruno

> * Bruno Haible <oehab@...> [2005-02-25 16:36:09 +0100]:
>
> Sam wrote:
>> > [2]> (language-information)
>> >
>> > *** - invalid byte sequence #xCF #xED in CHARSET:UTF-8 conversion
>> > The following restarts are available:
>> > ABORT          :R1      ABORT
>> > Break 1 [3]>
>> >
>> > clisp option "-E xxx" where xxx is CP1251, CP866 or UTF-8 does not
>> > influence result
>
> The macro res_to_obj() in modules/i18n/gettext.c uses UTF-8 always. I think
> it should better use *MISC-ENCODING* instead.

UTF-8 is hard-wired in modules/i18n/gettext.c in many many places.
when should it be replaced with *MISC-ENCODING*?

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://pmw.org.il/> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.honestreporting.com>
MS Windows: error: the operation completed successfully.

Sam Steingold:

> > [2]> (language-information)
> >
> > *** - invalid byte sequence #xCF #xED in CHARSET:UTF-8 conversion
> > The following restarts are available:
> > ABORT          :R1      ABORT
> > Break 1 [3]>
> >
> > clisp option "-E xxx" where xxx is CP1251, CP866 or UTF-8 does not
> > influence result
>
> apparently the return value of language-information contains some bytes
> which are not representable in the encodings you tried.
> you need a 1:1 encoding.
> how about "clisp -E iso-8859-1"?
>
iso-8859-[1..5] - no effect

> UTF-8 is hard-wired in modules/i18n/gettext.c in many many places.
> when should it be replaced with *MISC-ENCODING*?
>
How do it?

-- 
WBR, Yaroslav Kavenchuk.

> * Yaroslav Kavenchuk <xnirapuhx@...> [2005-02-28 17:20:34 +0200]:
>
>> UTF-8 is hard-wired in modules/i18n/gettext.c in many many places.
>> when should it be replaced with *MISC-ENCODING*?
>>
> How do it?

done - it's in the CVS now

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.honestreporting.com> <http://www.iris.org.il> <http://pmw.org.il/>
<http://www.mideasttruth.com/> <http://www.memri.org/> <http://www.camera.org>
MS Windows vs IBM OS/2: Why marketing matters more than technology...