develooper Front page | perl.perl5.porters | Postings from April 2013

Re: Blead on s390x

Thread Previous | Thread Next
From:
H.Merijn Brand
Date:
April 7, 2013 10:24
Subject:
Re: Blead on s390x
Message ID:
20130407090037.2c3e48bf@pc09.procura.nl
On Sat, 6 Apr 2013 22:05:16 +0200, Henri Kuiper
<henrikuiper@zdevops.com> wrote:

> Lol, just my 2cc bit might it be related to the same S**T I have with
> MongoDB? bigEndian vs SmallEndian?

I don't think so, as I smoke on PA-RISC (32bit and 64bit) daily, and
that should surface Endianess problems

On Sat, 6 Apr 2013 20:12:41 +0100, Nicholas Clark <nick@ccl4.org> wrote:

> On Sat, Apr 06, 2013 at 10:21:56AM +0200, H.Merijn Brand wrote:
> > I got a hercules iso from Henri yesterday on the OpenSUSE 12.3 launch
> > party. After converting the iso to something Linux understands (OS/X
> > iso's are not recognized as iso), it just took minutes to get it up
> > and running under hercules. It is not EBCDIC, but for sure it is
> > something different.
> 
> Is anyone actually using Perl on it?

As I understood, there is an increase of Linux use on mainframes, Henri
should be able to show more precise numbers.

> In that, we've not had any bug reports about this before, and for a lot of
> these somewhat esoteric platforms I'm sort of wondering at what point do
> they stop being "fun", and start being "work". Right now, I think it's more
> at the "fun" level, and it might be telling us something interesting about
> portability, as it might be that all these tests failing are down to the
> same problem.

At the moment, for me it is indeed withing the "fun" range, but with a
definite hint towards portability tests and sanity checks for
Configure. It already showed the use of good documentation of known
issues. I *remembered* the ODBM bug and swiftly found the solution in
our own docs.

When it starts being "work", I am unsure if the time it takes warrents
the effort.

FWIW I was amazed that it built till the end with so few fails

> Of all of these:
> 
> > op/pow.t                                                        (Wstat: 0 Tests: 267 Failed: 1)
> >   Failed test:  13
> > op/range.t                                                      (Wstat: 0 Tests: 141 Failed: 25)
> >   Failed tests:  84-95, 99-111
> > op/pack.t                                                       (Wstat: 0 Tests: 14704 Failed: 1)
> >   Failed test:  3499
> 
> pow.t looks to be the most interesting. What does *it* say. The test is:
> 
>     # This calculation ought to be within 0.001 of the right answer.
>     my $bits_in_uv = int (0.001 + log (~0+1) / log 2);

With the installed perl5.10

$ test@zdolinux:~> perl -wle'print int (0.001 + log (~0+1) / log 2)'
63

>     ...
>     
>     # Ought to be 32, 64, 36 or something like that.
>     
>     my $remainder = $bits_in_uv & 3;
>     
>     cmp_ok ($remainder, '==', 0, 'Sanity check bits in UV calculation')
>         or printf "# ~0 is %d (0x%d) which gives $bits_in_uv bits\n", ~0, ~0;
> 
> 
> > test@zdolinux:/perl/perl-git/t> ./perl harness op/numconvert.t
> > op/numconvert.t .. # '9.22337e+18' ne '1.84467e+19',    18446744073709551615   => P N N vs P N
> > # Failed test 15 - at op/numconvert.t line 247
> > # '1.84467e+19' ne '9.22337e+18',       -1      => U P N vs U N
> > # '1.84467e+19' ne '9.22337e+18',       18446744073709551615    => U P N vs U N
> 
> That's a factor of two out. Curious.
> 
> What's the output from
> 
>     perl -le 'printf "%X %X\n", ~0, -1'

test@zdolinux:~> perl -le 'printf "%X %X\n", ~0, -1'
FFFFFFFFFFFFFFFF FFFFFFFFFFFFFFFF

> Nicholas Clark

-- 
H.Merijn Brand  http://tux.nl   Perl Monger  http://amsterdam.pm.org/
using perl5.00307 .. 5.17   porting perl5 on HP-UX, AIX, and openSUSE
http://mirrors.develooper.com/hpux/        http://www.test-smoke.org/
http://qa.perl.org   http://www.goldmark.org/jeff/stupid-disclaimers/

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About