Bug #95239 for Net-SMTP-Bulk: Test suite fails with perl 5.19.x

Thu May 01 12:53:39 2014 SREZIC [...] cpan.org - Ticket created

Subject:

Test suite fails with perl 5.19.x

The pod-coverage.t test is failing for newer perls (5.19.x). See http://matrix.cpantesters.org/?dist=Net-SMTP-Bulk%200.16 for an overview. Regards, Slaven

Thu May 01 13:52:27 2014 ether [...] cpan.org - Correspondence added

This test should be moved to xt/ so it doesn't run for users, or guarded with: use Test::More; plan skip_all => 'these tests are for authors only' unless $ENV{RELEASE_TESTING};

Thu May 01 13:52:27 2014 The RT System itself - Status changed from 'new' to 'open'

Thu May 01 14:00:45 2014 SREZIC [...] cpan.org - Correspondence added

On 2014-05-01 13:52:27, ETHER wrote: Show quoted text

> This test should be moved to xt/ so it doesn't run for users, or > guarded with: > > use Test::More; > plan skip_all => 'these tests are for authors only' unless > $ENV{RELEASE_TESTING};

Actually it does not look like a pod coverage error at all. The same test runs fine with older perls < 5.19.0.

Sun May 04 13:55:09 2014 ANDK [...] cpan.org - Correspondence added

Show quoted text

> The same test runs fine with older perls < 5.19.0.

Not correct. There were these fails with 5.{10,12,14,16,18,19}.* http://analysis.cpantesters.org/reports_by_field?SUBMIT_xxx=Submit&distv=Net-SMTP-Bulk-0.16&field=fail%3At%2Fpod-coverage.t&field=meta%3Aperl

Sun May 04 14:28:47 2014 SREZIC [...] cpan.org - Correspondence added

On 2014-05-04 13:55:09, ANDK wrote: Show quoted text

> > The same test runs fine with older perls < 5.19.0.

> > Not correct. There were these fails with 5.{10,12,14,16,18,19}.* > > http://analysis.cpantesters.org/reports_by_field?SUBMIT_xxx=Submit&distv=Net- > SMTP-Bulk-0.16&field=fail%3At%2Fpod-coverage.t&field=meta%3Aperl

Linux is another story, but for FreeBSD there is a clear PASS/FAIL border between 5.18.2 and 5.19.7.

Fri May 09 22:08:20 2014 ANDK [...] cpan.org - Correspondence added

Show quoted text

> Linux is another story, but for FreeBSD there is a clear PASS/FAIL > border between 5.18.2 and 5.19.7.

Net-SMTP-Bulk recommends Coro and the pod-coverage error has something to do with Coro. I'd suggest you install Coro where you haven't and uninstall it where you have. That would test whether the clear border hypothesis holds.

Sat Aug 02 00:19:14 2014 knowzero [...] cpan.org - Correspondence added

On Sun May 04 14:28:47 2014, SREZIC wrote: Show quoted text

> On 2014-05-04 13:55:09, ANDK wrote:

> > > The same test runs fine with older perls < 5.19.0.

> > > > Not correct. There were these fails with 5.{10,12,14,16,18,19}.* > > > > http://analysis.cpantesters.org/reports_by_field?SUBMIT_xxx=Submit&distv=Net- > > SMTP-Bulk-0.16&field=fail%3At%2Fpod-coverage.t&field=meta%3Aperl

> > Linux is another story, but for FreeBSD there is a clear PASS/FAIL > border between 5.18.2 and 5.19.7.

Sorry for the delay in response. The problem might be related to use of Coro, that said, Coro should technically be optional. Initially the module was based on Coro but I implemented an upgrade in 0.16 to allow for AnyEvent only usage. So technically if you don't plan on using Coro and plan on using it in AnyEvent only mode, Coro should not be required. (By setting Mode => 'AnyEvent') I plan to do a few minor fixes for 0.17 and during that time I will see what I can do about getting the dependencies and tests worked out. I will probably also have to decide if to make AnyEvent mode the default (right now Coro is default). AnyEvent mode has less dependencies so I would prefer to make that default. But while the syntax of AnyEvent mode and Coro mode are the same, they behave differently. Once I figure out how I am going to proceed, I will release 0.17 with a fix to dependencies.

Sat Aug 30 04:36:01 2014 SREZIC [...] cpan.org - Correspondence added

On 2014-08-02 00:19:14, KnowZero wrote: Show quoted text

> On Sun May 04 14:28:47 2014, SREZIC wrote:

> > On 2014-05-04 13:55:09, ANDK wrote:

> > > > The same test runs fine with older perls < 5.19.0.

> > > > > > Not correct. There were these fails with 5.{10,12,14,16,18,19}.* > > > > > > http://analysis.cpantesters.org/reports_by_field?SUBMIT_xxx=Submit&distv=Net- > > > SMTP-Bulk-0.16&field=fail%3At%2Fpod-coverage.t&field=meta%3Aperl

> > > > Linux is another story, but for FreeBSD there is a clear PASS/FAIL > > border between 5.18.2 and 5.19.7.

> > > Sorry for the delay in response. The problem might be related to use > of Coro, that said, Coro should technically be optional. Initially the > module was based on Coro but I implemented an upgrade in 0.16 to allow > for AnyEvent only usage. So technically if you don't plan on using > Coro and plan on using it in AnyEvent only mode, Coro should not be > required. (By setting Mode => 'AnyEvent') > > I plan to do a few minor fixes for 0.17 and during that time I will > see what I can do about getting the dependencies and tests worked out. > > I will probably also have to decide if to make AnyEvent mode the > default (right now Coro is default). AnyEvent mode has less > dependencies so I would prefer to make that default. But while the > syntax of AnyEvent mode and Coro mode are the same, they behave > differently. > > Once I figure out how I am going to proceed, I will release 0.17 with > a fix to dependencies.

Maybe the best is to skip the pod coverage test if Coro is not available. Something like if (!eval { require Coro; 1 }) { plan skip_all => 'Coro is not installed, cannot test pod coverage'; exit } Or move the pod coverage tests into xt/, as ETHER suggested.

Sun Nov 09 00:44:59 2014 knowzero [...] cpan.org - Correspondence added

Ok, fixed in 0.18 along with a lot of major patches to AnyEvent.

Sun Nov 09 00:45:00 2014 knowzero [...] cpan.org - Status changed from 'open' to 'resolved'

Sun Nov 09 00:45:00 2014 knowzero [...] cpan.org - Fixed in 0.18 added