gmane.lisp.gcl.devel

up-to-date setup instructions for Win32 GCL

Donald Winiecki — 2012-05-25T18:53:44

Hi,

It's about a year late but included below is the most up-to-date setup
directions for Win32 GCL.  This includes setting up and using a CVS client,
and building Win32 GCL on a current WinXP, WinVista and Win7.

Next it's the regular docs.

Best,

_don



===============================================
BUILDING NATIVE WIN32 GNU COMMON LISP FROM CVS
===============================================

The preferred build host system for the Mingw32 compiler is MSYS.

I use gcc version 3.3.1 and binutils 2.14.90, but earlier versions
of gcc back to 2.95 are OK provided that you remove the
"-fno-zero-initialized-in-bss" flag in "h/mingw.defs" before running
"configure".

Note that gcc 3.3.3 and gcc 3.4.0 do NOT work; likewise binutils 2.13.90
and 2.15.90.

The working binutils version can be found at:

http://gd.tuwien.ac.at/gnu/mingw/binutils-2.14.90-20030612-1.tar.gz

===============================================
INSTALL AND CONFIGURE TORTOISE CVS
===============================================

Download Tor

Re: Compiler does not give error if T clause in caseconstruct is not last one

Camm Maguire — 2012-03-25T12:37:17

Greetings!

cvs -z9 -q diff -u -r Version_2_6_8pre:2012-03-20 -r Version_2_6_8pre:2012-03-24


Index: cmpnew/gcl_cmpif.lsp
===================================================================
RCS file: /sources/gcl/gcl/cmpnew/gcl_cmpif.lsp,v
retrieving revision 1.1.2.1.14.1
retrieving revision 1.1.2.1.14.2
diff -u -u -r1.1.2.1.14.1 -r1.1.2.1.14.2
--- cmpnew/gcl_cmpif.lsp21 Jun 2006 20:03:05 -00001.1.2.1.14.1
+++ cmpnew/gcl_cmpif.lsp23 Mar 2012 14:13:07 -00001.1.2.1.14.2
< at >< at > -330,32 +330,55 < at >< at >
   (when (endp args) (too-few-args 'case 1 0))
   (let* ((info (make-info))
          (key-form (c1expr* (car args) info))
-         (clauses nil))
+         clauses)
     (cond ((subtypep (info-type (second key-form)) 'fixnum)
-   (return-from c1case  (c1expr (convert-case-to-switch
- args default )))))
-    (dolist (clause (cdr args))
-      (cmpck (endp clause) "The CASE clause ~S is illegal." clause)
-      (case (car clause)
-            ((nil))
-            ((t otherwise)
-             (when default
-

Re: Compiler does not give error if T clause in case construct is not last one

Faheem Mitha — 2012-03-25T06:23:15


On Sat, 24 Mar 2012, Camm Maguire wrote:


Thanks for the information. I think this commit must be the one.

revision 1.1.2.1.14.2
date: 2012-03-23 19:43:07 +0530;  author: camm;  state: Exp;  lines: +48 
-25;  commitid: OhlbKUngXCha01Yv;
case default error checking

How do I get the diff? Sorry, I've never used cvs.

                                   Regards, Faheem

Re: Compiler does not give error if T clause in caseconstruct is not last one

Camm Maguire — 2012-03-24T12:12:15

Greetings!

export CVSROOT=:pserver:anonymous< at >cvs.sv.gnu.org:/sources/gcl
cvs -z9 -q co -d gcl-2.6.8pre -r Version_2_6_8pre gcl

Take care,

Faheem Mitha <faheem< at >faheem.info> writes:

Re: Compiler does not give error if T clause in case construct is not last one

Faheem Mitha — 2012-03-23T19:07:17


On Fri, 23 Mar 2012, Camm Maguire wrote:




Thanks for the reply. Can you point me to the fix? I browsed the CVS 
repository, but didn't see anything.

                                                          Regards, Faheem

Re: Compiler does not give error if T clause in caseconstruct is not last one

Camm Maguire — 2012-03-23T14:27:28

Greetings!

Faheem Mitha <faheem< at >faheem.info> writes:


This is fine, thanks!  (Debian bts is good too.)

I've committed a fix in the stable branch for the problem you report.
I'll be uploading this to Debian shortly.  Please check it out and let
me know if problems persist.  And thanks for the report!

Compiler does not give error if T clause in case construct is not last one

Faheem Mitha — 2012-03-22T20:54:43

Hi,

The bugs page at http://savannah.gnu.org/bugs/?group=gcl looks dead, so I'm 
writing to the developer mailing list. If there is a more suitable place to 
report bugs, let me know and I'll send it there.

Consider the following session. I believe this is wrong.

faheem< at >orwell:/usr/bin$ gcl
GCL (GNU Common Lisp)  2.6.7 CLtL1    Jan 20 2012 20:04:53
Source License: LGPL(gcl,gmp), GPL(unexec,bfd,xgcl)
Binary License:  GPL due to GPL'ed components: (XGCL READLINE UNEXEC)
Modifications of this banner must retain notice of a compatible license
Dedicated to the memory of W. Schelter

Use (help) to get some basic information on how to use GCL.
Temporary directory for compiler files set to /tmp/


X

)

FOO


"matches t"
"matches t"

See also https://bugs.launchpad.net/sbcl/+bug/959687 for context.

Please CC me on any reply. Thanks.

                                                  Regards, Faheem

Re: ghci on arm

Donald Winiecki — 2012-01-27T23:43:35

The binutils used with the Win32 builds (2.14.90-20030612) is covered under
LGPL 2.1

Best,

_don

~~~~~~~~~~~~~~~~~~~~~~~~~
Don Winiecki, Ed.D., Ph.D.
Professor
Boise State University, College of Engineering
Department of Instructional & Performance Technology
1910 University Drive, Boise, Idaho 83725-2070 USA
E-mail: dwiniecki< at >boisestate.edu
WWW: http://ipt.boisestate.edu
Telephone: (+01) 208 426 1899
Fax: (+01) 208 426 1970
~~~~~~~~~~~~~~~~~~~~~~~~d



On Fri, Jan 27, 2012 at 7:51 AM, Camm Maguire <camm< at >maguirefamily.org>wrote:

_______________________________________________
Gcl-devel mailing list
Gcl-devel< at >gnu.org
https://lists.gnu.org/mailman/listinfo/gcl-devel

Re: ghci on arm

Camm Maguire — 2012-01-27T14:51:49

Greetings!

In cvs tag Version_2_6_8pre, the file o/sfaslelf.c has the bulk.
Platform specific headers are selected at configure time and stored in
h/.

Take care,

Ben Gamari <bgamari.foss< at >gmail.com> writes:

Re: ghci on arm

Camm Maguire — 2012-01-27T14:58:23

Greetings!  GCL does use LGPL.   I'm unaware of a BSD variant.  What is
the license of binutils?

Take care,

Karel Gardas <karel.gardas< at >centrum.cz> writes:

Re: ghci on arm

Camm Maguire — 2012-01-26T03:12:13

Greetings!  GCL has a working native linker now on armel and armhf.  We
have a generic elf relocation file, with machine specific headers
implementing only those relocs output by gcc.  I can step you through
the code if interested.

For example, here is the core file for elf32_arm, which should be almost
readable: 

#define R_ARM_THM_CALL        10
#define R_ARM_CALL 28
#define R_ARM_V4BX 40
#define R_ARM_THM_MOVW_ABS_NC 47
#define R_ARM_THM_MOVW_ABS    48
    case R_ARM_THM_CALL: 
      s+=a; 
      if (ELF_ST_TYPE(sym->st_info)==STT_FUNC) s|=1; 
      s-=p+4; /*FIXME maybe drop 4 and add_val below*/
      s=((long)s>>1); 
      massert(!(abs(s)&0xffc00000));  
      store_val(where,MASK(11),s>>11); 
      store_val(where,MASK(11)<<16,(s&0x7ff)<<16); 
      break; 
    case R_ARM_THM_MOVW_ABS_NC:
      s+=a;
      if (ELF_ST_TYPE(sym->st_info)==STT_FUNC) s|=1;
      s&=0xffff;
      s=((s>>12)&0xf)|(((s>>11)&0x1)<<10)|((s&0xff)<<16)|(((s>>8)&0x7)<<28);
      add_vals(where,~0L,s);
      break;
    case R_AR

Re: maxima build fails using gcl

Camm Maguire — 2012-01-23T14:52:02

Greetings!

Karsten Trulsen <karstent< at >math.uio.no> writes:


OK this is a clear sign of a kernel memory protection or library address
mapping "feature".  In your log, brk is returning without error
indicating that the memory range should be within the .data segment of
the running process.  EACESS makes no sense after this, IMHO.  It might
be useful to fire up gcl and report the contents of /proc/<pid>/maps,
and perhaps also search your logs (under /var/log) for kernel exception
messages.  Perhaps ask with Red Hat?  The key observation is that
memory allocated with brk cannot be set with execute permissions beyond
a certain range (looks like 0x4000000 and above).

Re: maxima build fails using gcl

Camm Maguire — 2012-01-20T19:50:00

Greetings, and thanks!

The relevant bits are:

[pid 30971] mprotect(0xf0b000, 22028288, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xf0b000, 22028288, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xfd8000, 32767, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xfdf000, 16383, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xfcc000, 45055, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xfeb000, 20479, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xf7d000, 4095, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xf6c000, 8191, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xf75000, 8191, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xf8e000, 4095, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xffd000, 20479, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xfef000, 32767, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129] mprotect(0xf8f000, 8191, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
[pid 32129]

Re: axiom: please try stripping binary on powerpc again

Camm Maguire — 2012-01-17T18:33:35

Greetings!  I just did a by hand build of maxima without the strip in
sid, and all is well.  Will try removing on the next upload of
maxima/acl2/axiom/hol88.  Is there reason to believe that there was a
bug which has now been fixed?

Take care,

Jonathan Nieder <jrnieder< at >gmail.com> writes:

Re: ASDF 2.019 released

Camm Maguire — 2012-01-09T17:48:00

Greetings!  I must confess not having found time to investigate asdf.
If someone could give me a short list of test commands, I could run
these on all gcl platforms and report the results.  Is there anything
machine dependent in ASDF?  Should it not either fail or work at the
lisp level under gcl?

Take care,

David Billinghurst <dbmaxima< at >gmail.com> writes:

Re: ASDF 2.019 released

David Billinghurst — 2012-01-07T23:07:53

Can anyone confirm on which platforms, if any, that adsf works under 
gcl.  This question was raised on the maxima mailing list recently.

_______________________________________________
Gcl-devel mailing list
Gcl-devel< at >gnu.org
https://lists.gnu.org/mailman/listinfo/gcl-devel

Re: ASDF 2.019 released

Camm Maguire — 2012-01-09T17:48:00

Greetings!  I must confess not having found time to investigate asdf.
If someone could give me a short list of test commands, I could run
these on all gcl platforms and report the results.  Is there anything
machine dependent in ASDF?  Should it not either fail or work at the
lisp level under gcl?

Take care,

David Billinghurst <dbmaxima< at >gmail.com> writes:

Re: ASDF 2.019 released

David Billinghurst — 2012-01-07T23:07:53

Can anyone confirm on which platforms, if any, that adsf works under 
gcl.  This question was raised on the maxima mailing list recently.

_______________________________________________
Gcl-devel mailing list
Gcl-devel< at >gnu.org
https://lists.gnu.org/mailman/listinfo/gcl-devel

Re: ASDF 2.019 released

David Billinghurst — 2012-01-07T23:07:53

Can anyone confirm on which platforms, if any, that adsf works under 
gcl.  This question was raised on the maxima mailing list recently.

_______________________________________________
Gcl-devel mailing list
Gcl-devel< at >gnu.org
https://lists.gnu.org/mailman/listinfo/gcl-devel

Re: read-from-string bug

Camm Maguire — 2012-01-03T15:47:56

Greetings!  This looks right -- please try and let me know and I'll
commit.

Index: read.d
===================================================================
RCS file: /sources/gcl/gcl/o/read.d,v
retrieving revision 1.14.4.1.2.2.2.4.6.1.6.1.4.9
diff -u -u -r1.14.4.1.2.2.2.4.6.1.6.1.4.9 read.d
--- read.d27 Aug 2010 16:53:25 -00001.14.4.1.2.2.2.4.6.1.6.1.4.9
+++ read.d3 Jan 2012 15:46:22 -0000
< at >< at > -2057,17 +2057,23 < at >< at >
 goto READ;
 }
 }
-if (eof_errorp == Cnil && recursivep == Cnil)
-< at >(return eof_value)
-end_of_stream(strm);
+/* if (eof_errorp == Cnil && recursivep == Cnil) */
+/* < at >(return eof_value) */
+/* end_of_stream(strm); */
 
 READ:
 if (recursivep == Cnil)
 preserving_whitespace_flag = TRUE;
-if (recursivep == Cnil)
+        detect_eos_flag = TRUE;
+        if (recursivep == Cnil)
 x = read_object_non_recursive(strm);
 else
 x = read_object_recursive(strm);
+if (x == OBJNULL) {
+if (eof_errorp == Cnil && recursivep == Cnil)
+< at >(return eof_value)
+end_of_stream(strm);
+}

Re: GCL 2.6.8 on Mac intel, GCC 4.6.1

Camm Maguire — 2012-01-02T22:35:11

Greetings!  Just build with 4.6.2 on Debian without issue.  Does this
problem persist?

Take care,

daly< at >axiom-developer.org writes: