On Mon, 6 May 2002, Melvin Smith wrote:
> The broken 64bit systems seem to be related to compiling
> Parrot with a different wordsize than Perl5, since I'm currently
> getting byteorder from Perl5 Config.pm.
>
> I guess we will have to compute it ourselves after Parrot
> has picked what type to use for INTVAL/opcode_t
I think you're right in general.
However, in my case, I was using the defaults from perl5 for both iv and
byteorder, so byteorder was indeed computed for the correct iv.
That is, Parrot picked 'long long' for INTVAL and 'long long' for
opcode_t. Parrot correctly listed those as having byteorder 87654321,
since I was on a SPARC. Recomputing it to the same value won't change it.
[pause while I run cvs update and re-run tests]
Yup, about the same results as yesterday. I now only get 750 warnings,
but the tests still abort with
perl t/harness
t/op/basic..........# Failed test (t/op/basic.t at line 6)
# got: 'PackFile_unpack: Not a Parrot PackFile!
# Parrot VM: Can't unpack packfile t/op/basic1.pbc.
# '
# expected: ''
Finally, here's the output of
od -x t/opt/basic1.pbc
0000000 0800 0000 5041 4420 0706 0504 0302 0100
0000020 0000 0000 0131 55a1 0000 0000 5045 524c
0000040 0000 0000 0000 0000 0000 0000 0000 0008
0000060 0000 0000 0000 0000 0000 0000 0000 0010
0000100 0000 0000 0000 0001 0000 0000 0000 0000
0000120
--
Andy Dougherty [EMAIL PROTECTED]
Dept. of Physics
Lafayette College, Easton PA 18042
