C library installation request
cr...@animalhead.com
About a month ago I asked on the CPAN authors list why my module
IPC::MMA wasn't getting more test reports, despite the fact that I
had reworked its Makefile.PL to output a message about the need for a
required C library that I don't have the right to include with
IPC::MMA. David Golden sent a nice response:
> Most automated smoke testing will just quietly exit and continue.
> The testers will never see the prompt. This is generally how
> automated testers like it (though there are exceptions). Stopping
> at every Nth module to figure out how to resolve someone's custom
> dependency takes extra time that donating some CPU cycles to the
> testing effort does not.
>
> NOTE --> "exit 0" means that *no* report is sent. Not even an
> UNKNOWN. This is just like what happens with Perl prerequisites.
> If a prerequisite can't be satisfied, then no report is sent.
>
> I think the most effective thing that will lead to more test
> reports is asking people to install libmm. That's best done via
> personal appeal on the cpan-testers-discuss or on #cpantesters-
> discuss on irc.perl.org. That means work for someone who maintains
> multiple virtual machines, so ask nicely. ;-)
Asking nicely is something I don't have much experience with, but
this is my attempt.
The mm library, which is a non-CPAN prerequisite for IPC::MMA, is
quite easy to download and install from http://www.ossp.org/pkg/lib/mm/
Almost certainly, none of you will have any problem with downloading
and installing mm. mm only applies to Unix/Linux systems, so Windows-
based testers should ignore this message.
A few days from now I will upload a new IPC::MMA 0.58001, that will
be available to your test systems in the near future. The main
change is that it uses /dev/urandom rather than /dev/random for
testing, so that it should run its tests faster than previous versions.
So, might it be possible that you could please find time sometime in
your busy schedule to install the mm library on your test system(s)?
Historically only two testers have had mm on their systems, now 3
including David Golden.
One or two residual problems remain with the current 0.58, that may
or not be the fault of my module.
If you could please install mm, the test results would be very
helpful to provide additional data on these.
Thank you very much,
Craig MacKenna
Los Gatos, CA
David Golden
> The mm library, which is a non-CPAN prerequisite for IPC::MMA, is quite easy
> to download and install from http://www.ossp.org/pkg/lib/mm/
>
> Almost certainly, none of you will have any problem with downloading and
> installing mm. mm only applies to Unix/Linux systems, so Windows-based
> testers should ignore this message.
And, FYI, for debian/ubuntu systems:
$ sudo apt-get libmm-dev
On my ubuntu box, it's in the "universe" repo set.
Nit to Craig -- I sure hope the next version doesn't prompt me about
libmm being in /usr/local/lib anymore! :-)
-- David
cr...@animalhead.com
exactly :-) on your suggestions.
Craig
George Greer
> And, FYI, for debian/ubuntu systems:
>
> $ sudo apt-get libmm-dev
Woo, that reaches my lazy/easy threshold. Installed.
--
George Greer
Martin Evans
> On Tue, Nov 3, 2009 at 10:25 PM, <cr...@animalhead.com> wrote:
>> The mm library, which is a non-CPAN prerequisite for IPC::MMA, is quite easy
>> to download and install from http://www.ossp.org/pkg/lib/mm/
>>
>> Almost certainly, none of you will have any problem with downloading and
>> installing mm. mm only applies to Unix/Linux systems, so Windows-based
>> testers should ignore this message.
>
> And, FYI, for debian/ubuntu systems:
>
> $ sudo apt-get libmm-dev
sudo apt-get install libmm-dev
> On my ubuntu box, it's in the "universe" repo set.
>
> Nit to Craig -- I sure hope the next version doesn't prompt me about
> libmm being in /usr/local/lib anymore! :-)
>
> -- David
>
>
Martin
--
Martin J. Evans
Easysoft Limited
http://www.easysoft.com
Chris 'BinGOs' Williams
> Hello CPAN testers,
>
Helo.
> The mm library, which is a non-CPAN prerequisite for IPC::MMA, is
> quite easy to download and install from http://www.ossp.org/pkg/lib/mm/
>
Added so far to:
alphaserver Debian Linux using apt-get
Ubuntu Linux server using apt-get
FreeBSD 6.4 using ports, it is in devel/mm
FreeBSD 7.2 using ports, it is in devel/mm
Darwin, compiled and installed from source tarball
As it is in FreeBSD ports, I'm confident that it'll be in OpenBSD ports
and pkgsrc (used by NetBSD, Midnight BSD and Dragonfly BSD), but it
looks like it is painless to install from source.
I'll roll it out across my smoke boxen as I bring them online for testing.
I'll look forward to investigating your module further as I am
after a viable portable alterative to IPC::Shareable, which is anything but,
( viable and portable that is >:) )
Cheers,
--
Chris Williams
aka BinGOs
PGP ID 0x4658671F
http://www.gumbynet.org.uk
==========================
David Cantrell
> The mm library, which is a non-CPAN prerequisite for IPC::MMA, is quite
> easy to download and install from http://www.ossp.org/pkg/lib/mm/
$ sudo apt-get install libmm-dev
...
$ perl-5.6.2/bin/cpan IPC::MMA
...
Please enter path to libmm.so: [/usr/local/lib]
that's the wrong path (should be /usr/lib). I see that later in
Makefile.PL you run 'which mm-config' to see if you can find that, so I
suggest doing that earlier, and if you can find it, interrogate it for
the path to use.
The path is even wronger on my Solaris testing box - there it's
/usr/local/home/david/cpantesting/lib. But mm-config is still in the
path, cos /usr/local/home/david/cpantesting/bin is in the path.
Anyway, with the right path, on Linux:
* 5.6.2: fail, perl appeared to hang at the end of the tests
(or possibly while generating the test report)
* 5.8.9: t/5_hash appears to hang. aborted.
* 5.10.1: t/5_hash appears to hang. aborted.
Test results for Solaris and Irix may (or may not, if they hang too) be
forthcoming! If they don't show up on the CPAN-testers website in the
next few days, assume it didn't work.
--
David Cantrell | Cake Smuggler Extraordinaire
fdisk format reinstall, doo-dah, doo-dah;
fdisk format reinstall, it's the Windows way
David Cantrell
>> On Tue, Nov 3, 2009 at 10:25 PM, <cr...@animalhead.com> wrote:
>>> The mm library, which is a non-CPAN prerequisite for IPC::MMA, is quite easy
>>> to download and install from http://www.ossp.org/pkg/lib/mm/
Anyone know what the FreeBSD package is called, if there is one?
--
David Cantrell | Bourgeois reactionary pig
Compromise: n: lowering my standards so you can meet them
David Cantrell
> Martin Evans wrote:
>>> On Tue, Nov 3, 2009 at 10:25 PM, <cr...@animalhead.com> wrote:
>>>> The mm library, which is a non-CPAN prerequisite for IPC::MMA, is quite easy
>>>> to download and install from http://www.ossp.org/pkg/lib/mm/
>> sudo apt-get install libmm-dev
>
> Anyone know what the FreeBSD package is called, if there is one?
Never mind, I see that the lovely Mr. Bingos already told us
--
David Cantrell | Godless Liberal Elitist
Deck of Cards: $1.29.
"101 Solitaire Variations" book: $6.59.
Cheap replacement for the one thing Windows is good at: priceless
-- Shane Lazarus
David Cantrell
> Test results for Solaris and Irix may (or may not, if they hang too) be
> forthcoming! If they don't show up on the CPAN-testers website in the
> next few days, assume it didn't work.
The module passed on Solaris, failed on FreeBSD, and failed to even
build on Irix:
cc-1077 cc: ERROR File = mma_alloc.c, Line = 216
The indicated declaration has no storage class or type specifier.
inline int mma_alloc_mask(void) {return sizeof(union mem_chunk_mc_u)-1;}
^
inline is a C99-ism. SGI's compiler is C89 by default - you need to cc
-c99 to make it understand stuff like that (and I have no idea how to do
that when building XS, or even whether it's a good idea - the binaries
resulting might not be compatible with a perl built with plain old cc
for all I know).
As a workaround, you could do something like ...
#if defined(_SGIAPI) || defined( __sgi )
#define inline 1;
#endif
or whatever the syntax is for making inline into a null-op. Or you
could turn mma_alloc_mask into a macro so it gets inlined without the
inline keyword.
Note that if you release another version, I don't normally report build
failures. It can't be done reliably enough to automate. If you want me
to check that any new version builds, let me know and I'll check manually.
--
David Cantrell | top google result for "topless karaoke murders"
Immigration: making Britain great since AD43
cr...@animalhead.com
as you suggest, for quite some (unpublished) time. I was
waiting so that testers would have mm loaded, but there's
no right way to stage such events.
The "appears to hang" may be the result of using /dev/random
before it has enough randomness. 0.58001 uses /dev/urandom
to avoid such hangups. It seems that Linux takes
/dev/randomness more seriously than FreeBSD which I use.
Thanks,
Craig MacKenna
O. STeffen BEYer
for what it's worth, I have just installed "mm" on "my" machine (actually
one of Ralf Engelschall's machines, by happenstance - Ralf is the author of
"mm"!) running FreeBSD 7.2-stable and installed IPC::MMA successfully with
CPAN.pm (passed all tests). You should see the CPAN::Reporter report on
CPAN::Testers soon.
Hope this helps. :-)
Best regards,
Steffen Beyer
2009/11/4 <cr...@animalhead.com>
> Hello CPAN testers,
>
> About a month ago I asked on the CPAN authors list why my module IPC::MMA
> wasn't getting more test reports, despite the fact that I had reworked its
> Makefile.PL to output a message about the need for a required C library that
> I don't have the right to include with IPC::MMA. David Golden sent a nice
> response:
>
> Most automated smoke testing will just quietly exit and continue. The
>> testers will never see the prompt. This is generally how automated testers
>> like it (though there are exceptions). Stopping at every Nth module to
>> figure out how to resolve someone's custom dependency takes extra time that
>> donating some CPU cycles to the testing effort does not.
>>
>> NOTE --> "exit 0" means that *no* report is sent. Not even an UNKNOWN.
>> This is just like what happens with Perl prerequisites. If a prerequisite
>> can't be satisfied, then no report is sent.
>>
>> I think the most effective thing that will lead to more test reports is
>> asking people to install libmm. That's best done via personal appeal on the
>> cpan-testers-discuss or on #cpantesters-discuss on irc.perl.org. That