Re: Module::Build 0.30_01 - please test

2008-12-16 Thread Slaven Rezic
Ken Williams kena...@gmail.com writes:

 On Mon, Dec 15, 2008 at 3:03 PM, Slaven Rezic sla...@rezic.de wrote:
 I was just surprised to find version.pm installed at two
 locations (arch-dependent and arch-independent), but this seems to be
 unrelated to the changes in 0.30_01.

 That sounds weird - do you mean the actual version.pm, or
 M::B::Version.pm ?  M::B doesn't install the former.


I mean version.pm. In version.pm's Build.PL the function
have_c_compiler() is used to decide whether the version.pm's XS part
should be built and installed or not. Unfortunately have_c_compiler()
does not only check if a C compiler is available, it also checks if
ExtUtils::CBuilder is already installed. My assumption is that in the
smoker run version.pm was installed twice, first time without
ExtUtils::CBuilder available (and hence installed into the
arch-independent location) and the 2nd time with ExtUtils::CBuilder
installed in the meantime.

It makes me somewhat nervous to know that modules gets installed at
different locations (site vs. core, or arch-dependent vs.
arch-independent). One never knows which location will win.

Regards,
Slaven

-- 
Slaven Rezic - slaven at rezic dot de

tkrevdiff - graphical display of diffs between revisions (RCS, CVS or SVN)
http://ptktools.sourceforge.net/#tkrevdiff


Re: Module::Build 0.30_01 - please test

2008-12-16 Thread Ken Williams
On Mon, Dec 15, 2008 at 3:03 PM, Slaven Rezic sla...@rezic.de wrote:
 I was just surprised to find version.pm installed at two
 locations (arch-dependent and arch-independent), but this seems to be
 unrelated to the changes in 0.30_01.

That sounds weird - do you mean the actual version.pm, or
M::B::Version.pm ?  M::B doesn't install the former.

 -Ken


Re: Module::Build 0.30_01 - please test

2008-12-15 Thread Slaven Rezic
David Golden xda...@gmail.com writes:

 On Fri, Dec 12, 2008 at 8:58 AM, David Cantrell da...@cantrell.org.ukwrote:

 Hopefully I'll have the time to do all this myself in the few days
 before Christmas, but I'm sure it would be a good idea for at least one
 other person to do it.


 I have it running now -- though not quite so vanilla a set up.  Results
 probably around Monday.

I also had a smoke run on a FreeBSD machine running perl 5.8.8. I
found no surprises --- test failures and successes were more-or-less
consistent with MB 0.30 and 0.30_01, and the installed files looked
fine. I was just surprised to find version.pm installed at two
locations (arch-dependent and arch-independent), but this seems to be
unrelated to the changes in 0.30_01.

Regards,
Slaven

-- 
Slaven Rezic - slaven at rezic dot de

Berlin Perl Mongers - http://berlin.pm.org


Re: Module::Build 0.30_01 - please test

2008-12-12 Thread David Cantrell
[Added CC to cpan-testers-discuss]

On Thu, Dec 11, 2008 at 07:25:37PM -0800, Eric Wilhelm wrote:

 Testers, it would be great if you could run through a sampling of 
 distros out-of-band (to not annoy authors with our alpha bugs.)  David, 
 do you have a write-up from the procedure you used on this last time?
   http://www.nntp.perl.org/group/perl.qa/2008/09/msg11418.html

1  build a new perl
 sh Configure -Dprefix=$HOME/mbtestperl -de
 (IIRC I didn't bother with any combinations of threads, 64bit-ness
 etc, as those shouldn't affect M::B itself)
2  install the official version of M::B and CPAN::Reporter
3  take a backup in a tarball; also backup ~/.cpanreporter/reports-sent.db
4  delete ~/.cpanreporter/reports-sent.db, disable all email sending,
5  configure C::R to never ever need manual intervention
6  test a bunch of modules
 for i in `cat list_of_modules`; do
   path_to_perl/bin/perl -MCPAN -e install qw($i)
 done
7  move the new ~/.cpanreporter/reports-sent.db aside
8  delete that perl, restore from the tarball (this gets rid of any cruft
 left over from the previous round of testing)
9  install the dev version of M::B
10 repeat step 6
11 send Eric the old reports-sent.db and the new one generated in step
 10

Or you can do the two test runs in parallel with two identical
(apart from -Dprefix) builds of perl, if you set PERL_CPAN_REPORTERDIR
differently in the environment for the two smokers.

Hopefully I'll have the time to do all this myself in the few days
before Christmas, but I'm sure it would be a good idea for at least one
other person to do it.

-- 
David Cantrell | Enforcer, South London Linguistic Massive

What a lovely day!  Now watch me spoil it for you.


Re: Module::Build 0.30_01 - please test

2008-12-12 Thread David Golden
On Fri, Dec 12, 2008 at 8:58 AM, David Cantrell da...@cantrell.org.ukwrote:

 Hopefully I'll have the time to do all this myself in the few days
 before Christmas, but I'm sure it would be a good idea for at least one
 other person to do it.


I have it running now -- though not quite so vanilla a set up.  Results
probably around Monday.

David