#322 callback library fails on Mac OS 10.4.4

build problems
closed-out-of-date
Bruno Haible
ffcall (13)
5
2008-01-22
2006-02-02
Tom Emerson
No

Building CLISP 2.38, building on Mac OS 10.4.4 with GCC 4.0.0, the FFI
does not get configured because the callback library's unit tests fail: only
if these pass will the FFI be built.

Specifically the minitests fail:

Melchizedek (0)% ./minitests
void f(void):
void f(void):
int f(void):->99
int f(void):->99
int f(int):(1)->2
int f(int):(1)->2
int f(2*int):(1,2)->3
int f(2*int):(1,2)->3
int f(4*int):(1,2,3,4)->10
int f(4*int):(1,2,3,4)->10
int f(8*int):(1,2,3,4,5,6,7,8)->36
int f(8*int):(1,2,3,4,5,6,7,8)->36
int f(16*int):(1,2,3,4,5,6,7,8,9,11,12,13,14,15,16,17)->143
zsh: illegal hardware instruction ./minitests
Melchizedek (132)%

I have not yet debugged this further.

Discussion

  • Jörg Höhle
    Jörg Höhle
    2006-02-06

    • assigned_to: hoehle --> haible
     
  • Jörg Höhle
    Jörg Höhle
    2006-02-22

    Logged In: YES
    user_id=377168

    Could you provide more information about your hardware?
    I'm asking, because other people are using CLISP with OS X
    10.4 AFAIK (not that I can help much, as I have no Mac).

     
  • Tom Emerson
    Tom Emerson
    2006-02-22

    Logged In: YES
    user_id=37068

    Dual PowerMac G5, 2GHz, 64-bit CPUs. Mac OS X 10.4.4.

    CLisp on 10.4.x once the socket issue is patched, but you will not have FFI
    support on a G5. I will try this on my Powerbook which uses a 32-bit G4.

    I also tried with GCC 3.4.x with the same results.

     
  • Jörg Höhle
    Jörg Höhle
    2006-02-28

    • labels: 355966 --> ffcall
     
  • Jörg Höhle
    Jörg Höhle
    2006-02-28

    Logged In: YES
    user_id=377168

    I'm afraid I can't help you. Maybe you can figure out
    yourself because you have both a G4 and G5 and can compare
    them side by side (or compare their assembly
    specifications?).
    Sam Steingold would maybe say "PTC".
    BTW, the socket issue is fixed now in CVS.
    I think your test with gcc-3.4 exhibiting the same results
    as gcc-4.0.0 likely rules out a GCC bug.
    Would the minitest binary from the G4 run on the G5 or does
    it crash there? How do the 2 minitest binaries differ?

    Regards,
    Jörg Höhl

     
  • Jörg Höhle
    Jörg Höhle
    2006-04-24

    Logged In: YES
    user_id=377168

    CLISP in CVS contains changes to autoconf/m4/ffcall code as
    of 2006-04-22.
    Could you please retest using CVS?

     
  • Thomas Lazar
    Thomas Lazar
    2007-03-29

    Logged In: YES
    user_id=1756354
    Originator: NO

    Hi,

    i have the same problem on a Powerbook G4 500mhz with OS X 10.4.9. Tried to build it via pkgsrc lang/clisp.

    For more information look here please: http://www.netbsd.org/cgi-bin/query-pr-single.pl?number=36101

    blueAnt:~ ibn$ gcc --version
    powerpc-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 (Apple Computer, Inc. build 5247)
    Copyright (C) 2005 Free Software Foundation, Inc.

    Regards,
    Thomas Lazar

     
  • Sam Steingold
    Sam Steingold
    2008-01-22

    • status: open --> closed-out-of-date