Re: Testing Tapper::MCP

[Alexandr Ciornii]

Hello, Steffen

There are several problematic dependencies even on Linux:
Safe fails tests on perl 5.10.1.
https://rt.cpan.org/Ticket/Display.html?id=55115 . Upgrading
version.pm helps.
Crypt::DH tests are very slow without Math::BigInt::GMP or Math::BigInt::Pari.
Math::GMP requires libgmp.

On Debian installing libmath-gmp-perl, libmath-bigint-gmp-perl helps
with Crypt::DH and Math::GMP.

After installing these 3 modules, Tapper::MCP installs correctly.

On Windows there are more problems.

2012/6/11 Steffen Schwigon <steffen....@amd.com>:
> Hi!
>
> I am polishing our current "Tapper"[1] release 4.0 via CPAN Testers.

> One module, Tapper::MCP, is not tested, probably due to its long running
> test suite (~10min on modern average PC) which might hit timeouts in the
> Smokers.

> May I kindly ask some of you for testing it explicitely, if possible[2]?



--
Alexandr Ciornii, http://chorny.net

[Slaven Rezic]

Steffen Schwigon <steffen....@amd.com> writes:

> Hi!
>
> I am polishing our current "Tapper"[1] release 4.0 via CPAN Testers.
>

> As usual it's a great help. I found lots of micro issues, dependency
> bugs, and platform specific gotchas. Thanks!

>
> One module, Tapper::MCP, is not tested, probably due to its long running
> test suite (~10min on modern average PC) which might hit timeouts in the
> Smokers.
>

> However, this "MCP" is the heart of our automation layer and machine
> scheduler, and I would like to get a feeling of it's functionality from
> CPAN Testers.

>
> May I kindly ask some of you for testing it explicitely, if possible[2]?
>

> Is there something else I could do with such long running test suites?
> And is the long duration the root case at all?
>

Chance of all tests passing for Tapper-MCP's dependencies is:

1.8%

(According to
http://deps.cpantesters.org/?module=Tapper%3A%3AMCP&perl=any+version&os=any+OS )

--
Slaven Rezic - slaven <at> rezic <dot> de
BBBike - route planner for cyclists in Berlin
WWW version: http://www.bbbike.de
Perl/Tk version for Unix and Windows: http://bbbike.sourceforge.net

[Curtis Jewell]

On Mon, Jun 11, 2012, at 18:35, Alexandr Ciornii wrote:
> Hello, Steffen
>
> There are several problematic dependencies even on Linux:
> Safe fails tests on perl 5.10.1.
> https://rt.cpan.org/Ticket/Display.html?id=55115 . Upgrading
> version.pm helps.
> Crypt::DH tests are very slow without Math::BigInt::GMP or
> Math::BigInt::Pari.
> Math::GMP requires libgmp.
>
> On Debian installing libmath-gmp-perl, libmath-bigint-gmp-perl helps
> with Crypt::DH and Math::GMP.
>
> After installing these 3 modules, Tapper::MCP installs correctly.
>
> On Windows there are more problems.

Well... depends on the distribution of Perl for Windows.

I only know about Strawberry, which has libgmp included, (compiled
CPU-agnostic), but I know Math::Pari has compilation problems on
anything but 32-bit perl and Windows, as far as I recall.

But Math::BigInt::GMP is already there, and so is Crypt::DH.

--Curtis

> 2012/6/11 Steffen Schwigon <steffen....@amd.com>:
> > Hi!
> >
> > I am polishing our current "Tapper"[1] release 4.0 via CPAN Testers.
>
> > One module, Tapper::MCP, is not tested, probably due to its long running
> > test suite (~10min on modern average PC) which might hit timeouts in the
> > Smokers.
>
> > May I kindly ask some of you for testing it explicitely, if possible[2]?
>
>
>
> --
> Alexandr Ciornii, http://chorny.net

--
Curtis Jewell
swor...@curtisjewell.name

%DCL-E-MEM-BAD, bad memory
-VMS-F-PDGERS, pudding between the ears

[Chris 'BinGOs' Williams]

On Mon, Jun 11, 2012 at 01:59:25PM +0200, Steffen Schwigon wrote:
> Hi!
>
> I am polishing our current "Tapper"[1] release 4.0 via CPAN Testers.
>

> As usual it's a great help. I found lots of micro issues, dependency
> bugs, and platform specific gotchas. Thanks!
>

> One module, Tapper::MCP, is not tested, probably due to its long running
> test suite (~10min on modern average PC) which might hit timeouts in the
> Smokers.
>

> However, this "MCP" is the heart of our automation layer and machine
> scheduler, and I would like to get a feeling of it's functionality from
> CPAN Testers.
>

> May I kindly ask some of you for testing it explicitely, if possible[2]?
>

> Is there something else I could do with such long running test suites?
> And is the long duration the root case at all?

As Chorny has pointed out there is a buried dependency on Crypt::DH.

This might explain the killed jobs on my smokers:

[21:46] < nexus> Distribution: 'A/AM/AMD/Tapper-MCP-4.0.2.tar.gz' finished with status '9'
[01:09] < nexus> Distribution: 'A/AM/AMD/Tapper-MCP-4.0.2.tar.gz' finished with status '9'
[01:27] < fable> Distribution: 'A/AM/AMD/Tapper-MCP-4.0.2.tar.gz' finished with status '9'
[11:15] < nexus> Distribution: 'A/AM/AMD/Tapper-MCP-4.0.2.tar.gz' finished with status '9'
[13:46] < fremen64> Distribution: 'A/AM/AMD/Tapper-MCP-4.0.3.tar.gz' finished with status '9'

'9' being the exit code of the smoke job, meaning the job was killed with -9

Adding AAAA::Crypt::DH as a prereq will ensure that either Math::BigInt::GMP or Math::BigInt::Pari
are installed so that Crypt::DH will function with some sort of speed.

All my smokeboxen have libgmp installed and AAAA::Crypt::DH favours Math::BigInt::GMP over
Math::BigInt::Pari if the libgmp headers/libraries are available.

Cheers,

--
Chris Williams
aka BinGOs
PGP ID 0x4658671F
http://www.gumbynet.org.uk
==========================

[David Cantrell]

On Mon, Jun 11, 2012 at 09:48:43PM +0200, Slaven Rezic wrote:
> Steffen Schwigon <steffen....@amd.com> writes:
> > Is there something else I could do with such long running test suites?

Make sure that there's some output, or spit out a warning at the
beginning of the time-consuming test file, so that testers can see that
it hasn't just hung.

> Chance of all tests passing for Tapper-MCP's dependencies is:
>
> 1.8%
>
> (According to
> http://deps.cpantesters.org/?module=Tapper%3A%3AMCP&perl=any+version&os=any+OS )

If you specify 'any version' of perl then it assumes that you mean 5.005
for the purposes of figuring out what's in core, so many of those
dependencies won't actually be dependencies on a more modern perl, and
it will include test failures on very old perls too. That figure
doesn't mean very much anyway:
http://deps.cpantesters.org/static/overall-chance.html

Proc::ProcessTable is the biggest culprit, and is likely to be quite
platform-dependent. The next biggest is String::Diff, which you may be
able to get rid of.

Finally, the biggest cause of things not getting tested is if they
require some external setup (a database, for example, or an environment
variable) or there's a non-perl dependency such as on a library or an
application.

--
David Cantrell | A machine for turning tea into grumpiness

Us Germans take our humour very seriously
-- German cultural attache talking to the Today Programme,
about the German supposed lack of a sense of humour, 29 Aug 2001

[Aristotle Pagaltzis]

* Steffen Schwigon <steffen....@amd.com> [2012-06-11 14:00]:

> However, this "MCP" is the heart of our automation layer and machine
> scheduler

Is it trying to take over the Pentagon yet?

End of line,
--
Aristotle Pagaltzis // <http://plasmasturm.org/>