> Jonathan Worthington <[EMAIL PROTECTED]> wrote:
> > Don't tend to test it much on cygwin, but have done it now. Of possible
> > interest is this warning when running Configure.pl:-
>
> > Determining some sizes...
> > Hmm, I see your chosen INTVAL isn't the same size as your pointers.
>
> This doesn't really look good. What is the about of myconfig, especially
> that part about /Types/?
Here is the comlete myconfig (done with --jitcapable=0):-
Summary of my parrot 0.0.11.2 configuration:
configdate='Wed Oct 29 14:33:56 2003'
Platform:
osname=multi, archname=cygwin-multi-64int
jitcapable=0, jitarchname=nojit,
jitosname=nojit, jitcpuarch=i386
execcapable=0
perl=perl
Compiler:
cc='gcc', ccflags='',
Linker and Libraries:
ld='gcc', ldflags=' -s -L/usr/local/lib',
cc_ldflags='',
libs='-lcrypt -lgdbm_compat'
Dynamic Linking:
so='.so', ld_shared='-shared',
ld_shared_flags=''
Types:
iv=long long, intvalsize=8, intsize=4, opcode_t=long long,
opcode_t_size=8,
ptrsize=4, ptr_alignment=4 byteorder=12345678,
nv=double, numvalsize=8, doublesize=8
> Can you Configure it with intval=int?
Yeah, that gets rid of the warning from configure. We still get all the
"cast to pointer from integer of different size" warnings during make. The
results of make test after setting intval=int are rather... uh...
interesting?! though!
Failed Test Stat Wstat Total Fail Failed List of Failed
---------------------------------------------------------------------------
t/src/basic.t 1 256 3 1 33.33% 3
t/src/exit.t 1 256 3 1 33.33% 3
t/src/extend.t 12 3072 12 12 100.00% 1-12
t/src/hash.t 10 2560 12 10 83.33% 1 3-8 10-12
t/src/intlist.t 4 1024 4 4 100.00% 1-4
t/src/io.t 12 3072 16 12 75.00% 1-4 6-7 9-13 15
t/src/list.t 2 512 2 2 100.00% 1-2
t/src/sprintf.t 2 512 3 2 66.67% 2-3
3 subtests skipped.
Failed 8/9 test scripts, 11.11% okay. 44/59 subtests failed, 25.42% okay.
> > You have to configure with --jitcapable=0 under cygwin just as you do
with
> > nmake/cl (see thread NCI broken on Win32 for why).
>
> Still in need for a pre-processor output of jit_cpu.c to track this
> down. I did configure with execcapable=0 and didn't have any of these
> offending symbols - strange.
I have the output of gcc -E -I./include src/jit_cpu.c (I did it with gcc in
Cygwin, but the errors are talking about the same thing as the ones from cl
are). It came to > 800 KB so I've uploaded it to
http://www.jwcs.net/~jonathan/parrot/cpu_jit-E.txt. FYI,
$ cat cpu_jit-E.txt | grep RTYPE | wc -l
700
> > Tests 11 and 12, like you say, cause big problems.
>
> Could have been the missing prototype of the va_arg function.
Dan believes it's working again now; if I find otherwise I'll let you know.
Jonathan
