build on cygwin with clisp

57 views
Skip to first unread message

oldk1331

unread,
Oct 20, 2017, 8:35:31 AM10/20/17
to fricas-devel
Has anyone built FriCAS on cygwin recently?

I use fricas-1.3.2 tarball, with updated config.{guess,sub}.

I tried to build it with clisp, but that fails quickly
in src/lisp, because clisp fails at "save-core".

Prof. Dr. Johannes Grabmeier FW

unread,
Oct 20, 2017, 8:51:12 AM10/20/17
to fricas...@googlegroups.com

no I haven't, but I very much interested to have a working version for my students, many with Windows.


Am 20.10.17 um 14:35 schrieb oldk1331:
-- 
Mit freundlichen Grüßen

Johannes Grabmeier

Fraktionsvorsitzender 
FREIE WÄHLER, Stadtrat Deggendorf

Prof. Dr. Johannes Grabmeier
Köckstraße 1, D-94469 Deggendorf
Tel. +49-(0)-991-2979584, Tel. +49-(0)-151-681-70756
Fax: +49-(0)-322-241-92688

oldk1331

unread,
Oct 20, 2017, 9:28:34 AM10/20/17
to fricas-devel
On Fri, Oct 20, 2017 at 8:51 PM, Prof. Dr. Johannes Grabmeier FW
<joha...@grabmeier.net> wrote:
> no I haven't, but I very much interested to have a working version for my
> students, many with Windows.
>

What version will you prefer?

1. cygwin version requires to install cygwin. If you also install xming,
you can have hyperdoc. But this is built with clisp thus will be slow.
(and currently not working?)

2. mingw version with sbcl. I haven't tried to build it, but should be
easier than cygwin. It doesn't have hyperdoc and drawing.
But I think it can be used inside TeXmacs.

3. on Windows 10 and use Bash on Windows (WSL). This runs the
linux version binary. This is very fast but requires windows 10.
Hyperdoc should work too.

Grégory Vanuxem

unread,
Oct 20, 2017, 9:30:03 AM10/20/17
to fricas...@googlegroups.com
Hi,

I think cygwin version of CLisp no longer support dumping a core file.

To be investigated

--
Greg


--
You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fricas-devel+unsubscribe@googlegroups.com.
To post to this group, send email to fricas...@googlegroups.com.
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.

Grégory Vanuxem

unread,
Oct 20, 2017, 9:33:18 AM10/20/17
to fricas...@googlegroups.com
Hello,

I own a x86-64 SBCL version built on top of MinGW that hurts only one time in the test suite (quite an achievement :-)

But of course no X on Windows

Cheers

--
Greg

--

Grégory Vanuxem

unread,
Oct 20, 2017, 9:35:31 AM10/20/17
to fricas...@googlegroups.com
2017-10-20 15:28 GMT+02:00 oldk1331 <oldk...@gmail.com>:


3. on Windows 10 and use Bash on Windows (WSL).  This runs the
linux version binary.  This is very fast but requires windows 10.
Hyperdoc should work too.

Not aware of this, thanks for the information!
 


oldk1331

unread,
Oct 20, 2017, 10:44:25 AM10/20/17
to fricas-devel
On Fri, Oct 20, 2017 at 9:30 PM, Grégory Vanuxem <g.va...@gmail.com> wrote:
> Hi,
>
> I think cygwin version of CLisp no longer support dumping a core file.
>
> To be investigated
>

You mean recent version of clisp has this problem? So an ancient version
(clisp 2.49 in 2010) may work?

> Not aware of this, thanks for the information!

You can build fricas from source on WSL using SBCL 1.3.16~1.3.20.
Newer or older SBCL has problems.

oldk1331

unread,
Oct 20, 2017, 10:46:59 AM10/20/17
to fricas-devel
Hmm, it seems that SageMath is using ECL on cygwin.
ECL is not in cygwin repo, so you have to build it from source?
I tried but failed.

Grégory Vanuxem

unread,
Oct 20, 2017, 8:35:51 PM10/20/17
to fricas...@googlegroups.com
I'm not on a computer. Will talk you later.
It was clear. Let me know. 

Grégory Vanuxem

unread,
Oct 20, 2017, 8:38:01 PM10/20/17
to fricas...@googlegroups.com
Can CLisp dump a file???? 

oldk1331

unread,
Oct 26, 2017, 6:28:22 AM10/26/17
to fricas-devel
On Sat, Oct 21, 2017 at 8:37 AM, Grégory Vanuxem <g.va...@gmail.com> wrote:
> Can CLisp dump a file????

Clisp 2.49 on linux can. But that is released 7 years ago.

The cygwin version is a 2015 snapshot and it can't dump.
It gives an error message doesn't exist in version 2.49.

I can't build ECL-16.1.3 on cygwin either.

On the other hand, I don't think building on cygwin has any advantage.
It can provide user-unfriendly hyperdoc and graphics that Linux users
seldom use.

On that matter, mingw provides next to nothing: without hyperdoc and
graphics, only clef is lefted.

So, I'm thinking about a new packaging method for Windows:
distribute FriCAS like a Lisp application. In FriCAS release tarball,
there are already pre-generated Lisp files. I think it's possible
to package FriCAS as a Lisp program, just like Maxima.
One step further, maybe bundle TeXmacs with FriCAS.

Kurt Pagani

unread,
Oct 31, 2017, 9:37:06 PM10/31/17
to fricas-devel
It works on cygwin64 with ECL:

1. Open SageMath 8.0 Shell (run as Administrator)
2. Get and unpack FriCAS source tarball (e.g. 1.3.2)
3. patch configure configure.patch (use patch below)
4. ./configure --with-lisp=ecl --without-x
5. make & make install (install needs Admin rights)
6. fricas -->
    FriCAS Computer Algebra System
        Version: FriCAS 1.3.2
    Timestamp: Di, 31. Okt 2017 23:40:58

(sage-sh) kfp@NUC:~$ uname -a
CYGWIN_NT-6.1 NUC 2.8.2(0.313/5/3) 2017-07-12 10:58 x86_64 Cygwin

---

--- configure.orig    2017-04-25 13:41:49.000000000 +0200
+++ configure    2017-10-28 22:29:44.417592600 +0200
@@ -2302,6 +2302,7 @@ fi
 $as_echo "$ac_cv_build" >&6; }
 case $ac_cv_build in
 *-*-*) ;;
+*_*-*) ;;
 *) as_fn_error $? "invalid value of canonical build" "$LINENO" 5;;
 esac
 build=$ac_cv_build
@@ -2335,6 +2336,7 @@ fi
 $as_echo "$ac_cv_host" >&6; }
 case $ac_cv_host in
 *-*-*) ;;
+*_*-*) ;;
 *) as_fn_error $? "invalid value of canonical host" "$LINENO" 5;;
 esac
 host=$ac_cv_host
@@ -2368,6 +2370,7 @@ fi
 $as_echo "$ac_cv_target" >&6; }
 case $ac_cv_target in
 *-*-*) ;;
+*_*-*) ;;
 *) as_fn_error $? "invalid value of canonical target" "$LINENO" 5;;
 esac
 target=$ac_cv_target


Kurt Pagani

unread,
Nov 1, 2017, 12:14:59 AM11/1/17
to FriCAS - computer algebra system


On Thursday, 26 October 2017 12:28:22 UTC+2, oldk1331 wrote:
On Sat, Oct 21, 2017 at 8:37 AM, Grégory Vanuxem <g.va...@gmail.com> wrote:
> Can CLisp dump a file????

Clisp 2.49 on linux can.  But that is released 7 years ago.

The cygwin version is a 2015 snapshot and it can't dump.
It gives an error message doesn't exist in version 2.49.

I can't build ECL-16.1.3 on cygwin either.

I had no problems on x86_64 Cygwin (git from https://gitlab.com/embeddable-common-lisp/ecl.git)
$ ecl --version
ECL 16.1.3
 

On the other hand, I don't think building on cygwin has any advantage.
It can provide user-unfriendly hyperdoc and graphics that Linux users
seldom use.

Indeed :)
However, you can use it with texmacs or sage.
 

On that matter, mingw provides next to nothing: without hyperdoc and
graphics, only clef is lefted.

MSYS2/MingW64 looks promising.
 

So, I'm thinking about a new packaging method for Windows:
distribute FriCAS like a Lisp application.  In FriCAS release tarball,
there are already pre-generated Lisp files.  I think it's possible
to package FriCAS as a Lisp program, just like Maxima.
One step further, maybe bundle TeXmacs with FriCAS.

That's an idea. I also think this is feasible. I did some experiments and eventually managed to get it working with ABCL/JVM (a longstanding dream),
There are still some ceveats which might be overcome with the help of the few adepts ;)
 

oldk1331

unread,
Nov 1, 2017, 1:59:49 AM11/1/17
to fricas-devel
On Wed, Nov 1, 2017 at 12:14 PM, Kurt Pagani <nil...@gmail.com> wrote:
>> I can't build ECL-16.1.3 on cygwin either.
>
>
> I had no problems on x86_64 Cygwin (git from
> https://gitlab.com/embeddable-common-lisp/ecl.git)
> $ ecl --version
> ECL 16.1.3

Good to know, thanks.

>> So, I'm thinking about a new packaging method for Windows:
>> distribute FriCAS like a Lisp application. In FriCAS release tarball,
>> there are already pre-generated Lisp files. I think it's possible
>> to package FriCAS as a Lisp program, just like Maxima.
>> One step further, maybe bundle TeXmacs with FriCAS.
>
>
> That's an idea. I also think this is feasible. I did some experiments and
> eventually managed to get it working with ABCL/JVM (a longstanding dream),
> https://github.com/nilqed/fricas0/releases
> There are still some ceveats which might be overcome with the help of the
> few adepts ;)

Good job! I'll try it ASAP. It might make Android version of FriCAS
(using ECL)
available.

oldk1331

unread,
Nov 2, 2017, 10:15:38 PM11/2/17
to fricas-devel
>> That's an idea. I also think this is feasible. I did some experiments and
>> eventually managed to get it working with ABCL/JVM (a longstanding dream),
>> https://github.com/nilqed/fricas0/releases
>> There are still some ceveats which might be overcome with the help of the
>> few adepts ;)
>
> Good job! I'll try it ASAP. It might make Android version of FriCAS
> (using ECL)
> available.

I'm using sbcl-1.4.1, and I get error in the last step --
from the bottom of fricas.lisp, it calls (|spad|), then it complains
"The function COMMON-LISP-USER::PUT is undefined".
It should look up PUT from package BOOT.

BTW, the Android version of ECL aborts regularly, don't know why yet.

Kurt Pagani

unread,
Nov 2, 2017, 11:38:55 PM11/2/17
to fricas-devel

I'm using sbcl-1.4.1, and I get error in the last step --
from the bottom of fricas.lisp, it calls (|spad|), then it complains
"The function COMMON-LISP-USER::PUT is undefined".
It should look up PUT from package BOOT.

That's strange, indeed. Unfortunately, my SBCL on Ubuntu is quite old, but it works (just tried).
I got a package lock when using SBCL 1.3.18 on Windows/x_64. I'll try an update tomorrow (1.4.1 must be rather new?)

Have you any idea where the PUT could be? I only find:

$ grep "(put " interp/*.lisp
interp/daase.lisp:      (put con 'loaded c)

Did you encounter any package locks?

kfp@zulu:~/devel$ git clone https://github.com/nilqed/fricas0.git
Cloning into 'fricas0'...
remote: Counting objects: 1494, done.
remote: Compressing objects: 100% (579/579), done.
remote: Total 1494 (delta 912), reused 1494 (delta 912), pack-reused 0
Receiving objects: 100% (1494/1494), 7.78 MiB | 6.81 MiB/s, done.
Resolving deltas: 100% (912/912), done.
Checking connectivity... done.
kfp@zulu:~/devel$ cd fricas0
kfp@zulu:~/devel/fricas0$ ls
fricas0  mkfricas0.sh  README
kfp@zulu:~/devel/fricas0$ cd fricas0/
kfp@zulu:~/devel/fricas0/fricas0$ ls
algebra  compile-algebra.lisp  fricas.lisp  interp  interp.lisp  lisp  msgs
kfp@zulu:~/devel/fricas0/fricas0$ sbcl --load "fricas"
This is SBCL 1.3.0, an implementation of ANSI Common Lisp.
More information about SBCL is available at <http://www.sbcl.org/>.

SBCL is free software, provided as is, with absolutely no warranty.
It is mostly in the public domain; some portions are provided under
BSD-style licenses.  See the CREDITS and COPYING files in the
distribution for more information.

; file: /home/kfp/devel/fricas0/fricas0/lisp/fricas-package.lisp
; in: DEFMACRO IN-PACKAGE
;     (DEFMACRO FRICAS-LISP::IN-PACKAGE (PACKAGE &REST FRICAS-LISP::OPTIONS)
;       `(IN-PACKAGE ,PACKAGE))
; --> SB-C::NAMED-DS-BIND SB-INT:BINDING*
; ==>
;   (LET* ((#:G0
;           (SB-C::CHECK-DS-LIST/&REST (CDR #:EXPR) 1 1
;                                      '(# PACKAGE &REST FRICAS-LISP::OPTIONS)))
;          (PACKAGE (POP #:G0))
;          (FRICAS-LISP::OPTIONS #:G0))
;     (BLOCK FRICAS-LISP::IN-PACKAGE `(IN-PACKAGE ,PACKAGE)))
;
; caught STYLE-WARNING:
;   The variable OPTIONS is defined but never used.
;
; compilation unit finished
;   caught 1 STYLE-WARNING condition

; file: /home/kfp/devel/fricas0/fricas0/lisp/fricas-lisp.lisp
; in: PROGN (DEFVAR *SAVED-TERMINAL-IO* *TERMINAL-IO*)
;     (PROGN
;      (DEFVAR FRICAS-LISP::*SAVED-TERMINAL-IO* *TERMINAL-IO*)
;      (SETF *TERMINAL-IO*
;              (MAKE-TWO-WAY-STREAM *STANDARD-INPUT* *STANDARD-OUTPUT*))
;      (SETF *INVOKE-DEBUGGER-HOOK*
;              (LAMBDA (FRICAS-LISP::SB-COND FRICAS-LISP::SB-HOOK) (SETF #)))
;      (SETF *EVALUATOR-MODE* :INTERPRET))
;
; caught STYLE-WARNING:
;   The variable SB-COND is defined but never used.
;
; caught STYLE-WARNING:
;   The variable SB-HOOK is defined but never used.
;
; compilation unit finished
;   caught 2 STYLE-WARNING conditions
Warning: HyperTeX macro table not found
   Re-reading compress.daase   Re-reading interp.daase
   FriCAS initialization: interpreter
   FriCAS initialization: database
   FriCAS initialization: constructors
   FriCAS initialization: history
                       FriCAS Computer Algebra System
                            Version: COMMON-LISP
                   Timestamp: Tue Oct 31 21:55:28 CET 2017
-----------------------------------------------------------------------------
   Issue )copyright to view copyright notices.
   Issue )summary for a summary of useful system commands.
   Issue )quit to leave FriCAS and return to shell.
-----------------------------------------------------------------------------

   Re-reading category.daase
   Re-reading operation.daase
   Re-reading browse.daase
(1) ->



 

BTW, the Android version of ECL aborts regularly, don't know why yet.

Do you use the CL REPL? If so, which version: ECL 12.12.1? This is buggy!
I'm using ECL 13.5.1+armhf in GNURoot Debian (apt install ecl). There is also an SBCL (1.2.4).

oldk1331

unread,
Nov 3, 2017, 12:53:39 AM11/3/17
to fricas-devel
On Fri, Nov 3, 2017 at 11:38 AM, Kurt Pagani <nil...@gmail.com> wrote:
>
>
>> I'm using sbcl-1.4.1, and I get error in the last step --
>> from the bottom of fricas.lisp, it calls (|spad|), then it complains
>> "The function COMMON-LISP-USER::PUT is undefined".
>> It should look up PUT from package BOOT.
>
>
> That's strange, indeed. Unfortunately, my SBCL on Ubuntu is quite old, but
> it works (just tried).
> I got a package lock when using SBCL 1.3.18 on Windows/x_64. I'll try an
> update tomorrow (1.4.1 must be rather new?)
>
> Have you any idea where the PUT could be? I only find:

It's in vmlisp.lisp

> $ grep "(put " interp/*.lisp
> interp/daase.lisp: (put con 'loaded c)
>
> Did you encounter any package locks?

Yes, that's the problem with newer SBCL, and resolved in newer FriCAS,
see SVN r2167. It seems that you are using FriCAS 1.3.0? Why not
update to 1.3.2?

Also, it doesn't work correctly with clisp-2.49. It gives error at
*** - EVAL: undefined function STTOSEX
which is a function defined in boot/ptyout.boot


>> BTW, the Android version of ECL aborts regularly, don't know why yet.
>
>
> Do you use the CL REPL? If so, which version: ECL 12.12.1? This is buggy!
> I'm using ECL 13.5.1+armhf in GNURoot Debian (apt install ecl). There is
> also an SBCL (1.2.4).
> https://play.google.com/store/apps/details?id=com.gnuroot.debian&hl=en

No, I cross-compiled ECL from source (just like MaximaOnAndroid) and use
it under Termux. Have you tried runnig FriCAS in GNURoot on your phone?

BTW, SBCL doesn't support PIE and can't run after Android 5.0. What's your
Android version?

Kurt Pagani

unread,
Nov 3, 2017, 12:02:09 PM11/3/17
to fricas-devel
On 3 November 2017 at 05:53, oldk1331 <oldk...@gmail.com> wrote:
On Fri, Nov 3, 2017 at 11:38 AM, Kurt Pagani <nil...@gmail.com> wrote:
>
>
>> I'm using sbcl-1.4.1, and I get error in the last step --
>> from the bottom of fricas.lisp, it calls (|spad|), then it complains
>> "The function COMMON-LISP-USER::PUT is undefined".
>> It should look up PUT from package BOOT.
>
>
> That's strange, indeed. Unfortunately, my SBCL on Ubuntu is quite old, but
> it works (just tried).
> I got a package lock when using SBCL 1.3.18 on Windows/x_64. I'll try an
> update tomorrow (1.4.1 must be rather new?)
>
> Have you any idea where the PUT could be? I only find:

It's in vmlisp.lisp

> $ grep "(put " interp/*.lisp
> interp/daase.lisp:      (put con 'loaded c)
>
> Did you encounter any package locks?

Yes, that's the problem with newer SBCL, and resolved in newer FriCAS,
see SVN r2167.  It seems that you are using FriCAS 1.3.0? Why not
update to 1.3.2?

Lack of time ;)
In the meantime I upgraded to SBCL 1.4.1 and Fricas 1.3.2:
SBCL complained about two missing files: i-util and dq ?
Otherwise it works: see log.txt

Summarized (what I've done):

1. wget  fricas/sbcl tarballs
2. configure/make
3. adjust mkfricas0.sh -> run -> creates fricas0
4. removed loading of i-util.lisp and dq.lisp
5. cd fricas0 -> sbcl --load "fricas.lisp"
...

I've pushed the new folder to github.
Ofr course, we have to check about dq and i-util.


Also, it doesn't work correctly with clisp-2.49.  It gives error at
  *** - EVAL: undefined function STTOSEX
which is a function defined in boot/ptyout.boot


May be connected to the old version. On the other hand, it might depend on by which lisp flavour the *.clisp files are generated.
I have no clear picture yet about this mechanisms.

 

>> BTW, the Android version of ECL aborts regularly, don't know why yet.
>
>
> Do you use the CL REPL? If so, which version: ECL 12.12.1? This is buggy!
> I'm using ECL 13.5.1+armhf in GNURoot Debian (apt install ecl). There is
> also an SBCL (1.2.4).
> https://play.google.com/store/apps/details?id=com.gnuroot.debian&hl=en

No, I cross-compiled ECL from source (just like MaximaOnAndroid) and use
it under Termux.  Have you tried runnig FriCAS in GNURoot on your phone?

No :) 
A dotty idea, but it ought work. I'll try on a tablet at least.

BTW, SBCL doesn't support PIE and can't run after Android 5.0.  What's your
Android version?


Version 7.0. It works fine in GNURoot.
log.txt

oldk1331

unread,
Nov 3, 2017, 8:54:07 PM11/3/17
to fricas-devel
Hi Kurt, mystery solved.

The problems happen at my ~/.fricas.input.
Your project doesn't include BOOT source code, so
)boot $IOindex := 0
will fail.

So maybe you should put following in README:

If there are any problems, try again with profile disabled:
export FRICAS_INITFILE=""


BTW, it also works with clisp-2.49. Thanks again.

Grégory Vanuxem

unread,
Nov 5, 2017, 12:36:41 AM11/5/17
to fricas...@googlegroups.com

 

 

De : Kurt Pagani
Envoyé le :mercredi 1 novembre 2017 05:15
À : FriCAS - computer algebra system
Objet :Re: [fricas-devel] build on cygwin with clisp

 



On Thursday, 26 October 2017 12:28:22 UTC+2, oldk1331 wrote:

On Sat, Oct 21, 2017 at 8:37 AM, Grégory Vanuxem <g.va...@gmail.com> wrote:
> Can CLisp dump a file????

Clisp 2.49 on linux can.  But that is released 7 years ago.

The cygwin version is a 2015 snapshot and it can't dump.
It gives an error message doesn't exist in version 2.49.

I can't build ECL-16.1.3 on cygwin either.

 

I had no problems on x86_64 Cygwin (git from https://gitlab.com/embeddable-common-lisp/ecl.git)
$ ecl --version
ECL 16.1.3

 


On the other hand, I don't think building on cygwin has any advantage.
It can provide user-unfriendly hyperdoc and graphics that Linux users
seldom use.

 

Indeed :)

However, you can use it with texmacs or sage.

 


On that matter, mingw provides next to nothing: without hyperdoc and
graphics, only clef is lefted.

 

MSYS2/MingW64 looks promising.

 

I agree  but too  buggy actually

 

We  must get rid of Lisp.

Simple to say

 


So, I'm thinking about a new packaging method for Windows:
distribute FriCAS like a Lisp application.  In FriCAS release tarball,
there are already pre-generated Lisp files.  I think it's possible
to package FriCAS as a Lisp program, just like Maxima.
One step further, maybe bundle TeXmacs with FriCAS.

 

That's an idea. I also think this is feasible. I did some experiments and eventually managed to get it working with ABCL/JVM (a longstanding dream),

There are still some ceveats which might be overcome with the help of the few adepts ;)

 

--

You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fricas-devel...@googlegroups.com.

Kurt Pagani

unread,
Nov 15, 2017, 12:41:25 AM11/15/17
to FriCAS - computer algebra system

No, I cross-compiled ECL from source (just like MaximaOnAndroid) and use
it under Termux.  Have you tried runnig FriCAS in GNURoot on your phone?

No :) 
A dotty idea, but it ought work. I'll try on a tablet at least.


Amazing. It actually works on the tablet as well as on the phone ;)
Needless to say, the compilation of the algebra took far more than an hour.
So ,I guess it will also do using Termux !?

I have simplified the lisp build and tested it on several platforms:

It also works in sage (using the built in ECL, see the .html), so that some marshalling interface should be feasible. This, however, would require Fricas providing some stack based procedural hook, e.g. like that of REDUCE (I've done it with sympy <=> REDUCE,  it's fairly straightforward), or some other method to get the internals in a simple way. Of course, eventually the interpreter can be compiled and packed into a faslib. It would be nice to have the algebra in a lib as well ;)

Do you think export FRICAS_INITFILE=""  is still necessary?
I've set $IOndex in ~/.fricas.input without problems.




sage_fricas0.html

oldk1331

unread,
Nov 15, 2017, 2:27:16 AM11/15/17
to fricas-devel
> Do you think export FRICAS_INITFILE="" is still necessary?
> I've set $IOndex in ~/.fricas.input without problems.

Yes, a single line ")boot $IOindex := 0" in ~/.fricas.input
will cause the following error in SBCL:

>> System error:
The function BOOTTRAN::STTOSEX is undefined.

After all, you didn't include the BOOT sources.

Kurt Pagani

unread,
Nov 15, 2017, 6:11:12 PM11/15/17
to fricas-devel
You were right, I could reproduce it on Linux (on cyg/win it's slightly different).
Added to readme.
BTW on arm the normal build work as well (see arm.txt). Fricas runs faster than I expected ;)



arm.txt

Qian Yun

unread,
Mar 10, 2021, 11:03:32 AM3/10/21
to fricas-devel
Late reply to myself after 3.5 years:

I hit this problem again, and the culprit is

clisp-2.49.92-1.src/cygwin-no-save-executable.patch

(from
https://cygwin.com/cgi-bin2/package-cat.cgi?file=x86_64%2Fclisp-src%2Fclisp-2.49.92-1-src&grep=clisp
)

which seems to be committed since April 2015.

- Qian
Reply all
Reply to author
Forward
0 new messages