Re: perl problems

[heymanj@acm.org]

On 9 May 2008 at 12:21, Stephen R Marenka <stephen@marenka.net> wrote:

> 
> So we haven't managed to build any version of perl > 5.10.0
> 
> This was built under aranym, if that matters. If ya'll think it's worth
> it, I can copy it to a non-aranym box and run the tests there (or even
> build it from scratch).
> 
> I was looking at the test failures and here are the relevant portions.
> 
> t/op/cmp.t
> | not ok 1471 # (NaN <=> inf) gives: undef $i=3 $j=12, < disagrees
> | not ok 1476 # (NaN <=> inf) gives: undef $i=3 $j=12, <= disagrees
> | not ok 1487 # (NaN <=> -inf) gives: undef $i=3 $j=13, > disagrees
> | not ok 1488 # (NaN <=> -inf) gives: undef $i=3 $j=13, >= disagrees
> 
> 
> t/op/int.t
> | not ok 8 # int(4294967303.15) is -2147483648, not 4294967303
> | not ok 9 # int(4294967303.15) is -2147483648, not -4294967303
> | not ok 11 # int(4294967295.7) is -2147483648, not 4294967295
> | not ok 12 # int(4294967296.11312) is -2147483648, not 4294967296
> | not ok 13 # int(279964589018079/59) is -2147483648, not 4745162525730
> | not ok 14 # int(279964589018079/59) is -2147483648, not 4745162525730
> 
> t/op/pack.t
> total wreck with 303 not ok lines
> 
> t/op/threads.t
> hangs forever with
> | 1..10
> | ok 1 - fresh_perl - delete() under threads
> | ok 2 - fresh_perl - weaken ref under threads
> | ok 3 - fresh_perl - weaken ref \#2 under threads
> | # parent 1253: continue
> | # kid 1 before sort
> | # kid 2 before sort
> | # parent 1253: continue
> | # parent 1253: waiting for join
> 
> The last is what finally kills the buildds.
> 
> Can anyone help out with this?

OK.  Just completed building/running tests on my A3000 w/A3640 and
131MB of memory.  Slow process to do it natively.  20+ hrs to build,
36 hrs for testing.

I took all the defaults during the 'Configure' interactions.

Running on sarge (v3r1):
   - 2.4.27-amiga kernel
   - gcc 3.4

All tests run to completion w/o any errors.  Am looking to upgrade
my system to etch-m68k in the near future.  Will retry after doing
the upgrade.

Note that since we're good at sarge (v3r1), something new was introduced
that would cause the problems you're seeing.  The combination of 2.6
kernel/glibc/gcc4 would be where the problem was introduced...

> Thanks,
> 
> Stephen
> 
> --
> Stephen R. Marenka     If life's not fun, you're not doing it right!
> <stephen@marenka.net>

jerry
       //  Jerry Heyman      | "Software is the difference between
      //   Amiga Forever :-) |  hardware and reality"
  \\ //    heymanj@acm.org   | 
   \X/     http://www.hobbeshollow.com