[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

FFI vacall fails check.



Hi. I'm new to this list, but have checked the mailing-list-archive
and can't see any report of this problem:

I unpacked the clisp-source (1996-07-22 from linux developers resource
sept '96 edition) and did:

./configure i386-unknown-linux

At the end of the configuration:
  [...]
  gcc -O  -I. -I../../ffcall/vacall -c ../../ffcall/vacall/tests.c
  gcc -O  tests.o vacall.o misc.o structcpy.o -o tests
  ./tests > tests.out
  uniq -u < tests.out > tests.output.i386-unknown-linux
  test '!' -s tests.output.i386-unknown-linux
  make: *** [check] Error 1

  To continue building CLISP, the following commands are recommended
  (cf. unix/INSTALL step 4):
      cd i386-unknown-linux
      ./makemake  --with-readline > Makefile
      make
      make check


The contents of the file
i386-unknown-linux/vacall/tests.output.i386-unknown-linux:
------------start--------- 
X f(B,char,double,B):({0.1,(1,2,3}},'^?',0.3,{0.2,{5,4,3}})->{"return val",'^?'}
X f(B,char,double,B):({0.1,{1,2,3}},'^?',0.3,{0.2,{5,4,3}})->{"^?",'^?'}
------------end-----------

(The '^?' is char 0x7f)

I can build clisp using ./makemake --with-readline etc., it passes all
the checks and runs perfectly. However I would like to build with the
--dynamic-ffi package, to interface with C code.
(The avcall and trampoline parts of ffcall compile and check OK, the
problem is just with vacall.)

I'm using:
 linux-2.0.12
 gcc-2.7.2 (with patch for strength-reduction bug)
 binutils-2.6.0.14
 libc-5.2.18

I'd appreciate hearing from anyone who has built the FFI package
without this problem, and what kernel/gcc/binutils version you used.

Thanks,

    Geoff Reynolds <geoff@codeworks.gen.nz>

--
geoff@codeworks.gen.nz (Geoff Reynolds)
The BBS Works -- +64 9 625 2851 NZL New Zealand's Oldest BBS !