Bug #127463 for SVG-Calendar: Can't locate object method "new" via package "Astro::Coord::ECI::Sun"

Fri Oct 26 13:52:47 2018 SREZIC [...] cpan.org - Ticket created

Subject:

Can't locate object method "new" via package "Astro::Coord::ECI::Sun"

The test suite started to fail on my smokers: ... Can't locate object method "new" via package "Astro::Coord::ECI::Sun" (perhaps you forgot to load "Astro::Coord::ECI::Sun"?) at /usr/perl5.28.0-RC2p/lib/site_perl/5.28.0/Astro/Coord/ECI.pm line 1605, <DATA> line 129. # Looks like your test exited with 2 before it could output anything. t/moon.t .......... Dubious, test returned 2 (wstat 512, 0x200) Failed 93/93 subtests ...

Fri Oct 26 13:57:59 2018 SREZIC [...] cpan.org - Correspondence added

RT-Send-CC:

wyant [...] cpan.org

On 2018-10-26 13:52:47, SREZIC wrote: Show quoted text

> The test suite started to fail on my smokers: > > ... > Can't locate object method "new" via package "Astro::Coord::ECI::Sun" > (perhaps you forgot to load "Astro::Coord::ECI::Sun"?) at > /usr/perl5.28.0-RC2p/lib/site_perl/5.28.0/Astro/Coord/ECI.pm line > 1605, <DATA> line 129. > # Looks like your test exited with 2 before it could output anything. > t/moon.t .......... > Dubious, test returned 2 (wstat 512, 0x200) > Failed 93/93 subtests > ...

Maybe this happens if newer versions of Astro-satpass are installed.

Sat Oct 27 10:23:12 2018 wyant [...] cpan.org - Correspondence added

Thanks for adding me to this ticket. I will take a look. Tom

Sat Oct 27 10:23:13 2018 The RT System itself - Status changed from 'new' to 'open'

Sat Oct 27 11:57:30 2018 wyant [...] cpan.org - Correspondence added

At this point this looks like my bug. Astro::Coord::ECI::Moon 0.100 was the last good version. In 0.100_01 I put in support for alternate Sun classes (in preparation for the release of Astro::Coord::ECI::VSOP87D), and in that change 'use Astro::Coord::ECI::Sun' was removed from Astro::Coord::ECI::Moon and Astro::Coord::ECI::Star.

Sat Oct 27 12:29:54 2018 wyant [...] cpan.org - Reference by ticket #127488 added

Sat Oct 27 12:30:48 2018 wyant [...] cpan.org - Correspondence added

I have opened RT 127488 against Astro-satpass

Sat Oct 27 12:51:32 2018 wyant [...] cpan.org - Correspondence added

Astro-satpass version 0.102_01, which I believe and hope fixes this, just went to PAUSE. FWIW, the minimum code to reproduce the problem (or not) is: use Astro::Coord::ECI; Astro::Coord::ECI->new()->get( 'sun' );

Sat Nov 03 11:30:30 2018 wyant [...] cpan.org - Correspondence added

On Sat Oct 27 12:51:32 2018, WYANT wrote: Show quoted text

> Astro-satpass version 0.102_01, which I believe and hope fixes this, > just went to PAUSE. FWIW, the minimum code to reproduce the problem > (or not) is: > > use Astro::Coord::ECI; > Astro::Coord::ECI->new()->get( 'sun' );

No test failures in Astro-satpass 0.102_01, versus 473 passes. I have published Astro-satpass 0.103, and closed RT 127488. I suspect this ticket can be closed also.

Sat Nov 03 12:50:47 2018 SREZIC [...] cpan.org - Correspondence added

RT-Send-CC:

wyant [...] cpan.org

On 2018-11-03 11:30:30, WYANT wrote: Show quoted text

> On Sat Oct 27 12:51:32 2018, WYANT wrote:

> > Astro-satpass version 0.102_01, which I believe and hope fixes this, > > just went to PAUSE. FWIW, the minimum code to reproduce the problem > > (or not) is: > > > > use Astro::Coord::ECI; > > Astro::Coord::ECI->new()->get( 'sun' );

> > > No test failures in Astro-satpass 0.102_01, versus 473 passes. I have > published Astro-satpass 0.103, and closed RT 127488. I suspect this > ticket can be closed also.

A possible action would be to adjust the minimum prereq version in this distribution, so I'll leave the decision about closing this ticket to the author. But I can confirm that the newer version indeed fixes the problem.

Sun Nov 04 21:52:41 2018 IVANWILLS [...] cpan.org - Correspondence added

I've increased the minimum version and released v0.3.13 On Sun Nov 04 03:50:47 2018, SREZIC wrote: Show quoted text

> On 2018-11-03 11:30:30, WYANT wrote:

> > On Sat Oct 27 12:51:32 2018, WYANT wrote:

> > > Astro-satpass version 0.102_01, which I believe and hope fixes > > > this, > > > just went to PAUSE. FWIW, the minimum code to reproduce the problem > > > (or not) is: > > > > > > use Astro::Coord::ECI; > > > Astro::Coord::ECI->new()->get( 'sun' );

> > > > > > No test failures in Astro-satpass 0.102_01, versus 473 passes. I have > > published Astro-satpass 0.103, and closed RT 127488. I suspect this > > ticket can be closed also.

> > A possible action would be to adjust the minimum prereq version in > this distribution, so I'll leave the decision about closing this > ticket to the author. > > But I can confirm that the newer version indeed fixes the problem.

Sun Nov 04 21:52:42 2018 IVANWILLS [...] cpan.org - Status changed from 'open' to 'resolved'

Sun Nov 04 21:52:42 2018 IVANWILLS [...] cpan.org - Taken

Sun Nov 04 21:52:42 2018 IVANWILLS [...] cpan.org - Fixed in v0.3.13 added

Mon Nov 05 01:13:46 2018 SREZIC [...] cpan.org - Correspondence added

On 2018-11-04 21:52:41, IVANWILLS wrote: Show quoted text

> I've increased the minimum version and released v0.3.13 > > On Sun Nov 04 03:50:47 2018, SREZIC wrote:

> > On 2018-11-03 11:30:30, WYANT wrote:

> > > On Sat Oct 27 12:51:32 2018, WYANT wrote:

> > > > Astro-satpass version 0.102_01, which I believe and hope fixes > > > > this, > > > > just went to PAUSE. FWIW, the minimum code to reproduce the problem > > > > (or not) is: > > > > > > > > use Astro::Coord::ECI; > > > > Astro::Coord::ECI->new()->get( 'sun' );

> > > > > > > > > No test failures in Astro-satpass 0.102_01, versus 473 passes. I have > > > published Astro-satpass 0.103, and closed RT 127488. I suspect this > > > ticket can be closed also.

> > > > A possible action would be to adjust the minimum prereq version in > > this distribution, so I'll leave the decision about closing this > > ticket to the author. > > > > But I can confirm that the newer version indeed fixes the problem.

> >

I still see failures if an older Astro::satpass is installed. Unfortunately it seems that setting a higher version for a recommended module does not force a reinstall. Maybe the only solution is to also check the version in the test script, and skip if it's the problematic version.

Mon Nov 05 01:13:47 2018 SREZIC [...] cpan.org - Status changed from 'resolved' to 'open'

Mon Nov 05 01:13:47 2018 SREZIC [...] cpan.org - Fixed in v0.3.13 deleted

Tue Jan 29 15:28:14 2019 SREZIC [...] cpan.org - Broken in v0.3.13 added

Tue Jan 29 15:28:14 2019 SREZIC [...] cpan.org - Broken in v0.3.14 added

Tue Jan 29 15:36:31 2019 SREZIC [...] cpan.org - Correspondence added

RT-Send-CC:

ANDK [...] cpan.org

On 2018-11-05 01:13:46, SREZIC wrote: Show quoted text

> On 2018-11-04 21:52:41, IVANWILLS wrote:

> > I've increased the minimum version and released v0.3.13 > > > > On Sun Nov 04 03:50:47 2018, SREZIC wrote:

> > > On 2018-11-03 11:30:30, WYANT wrote:

> > > > On Sat Oct 27 12:51:32 2018, WYANT wrote:

> > > > > Astro-satpass version 0.102_01, which I believe and hope fixes > > > > > this, > > > > > just went to PAUSE. FWIW, the minimum code to reproduce the > > > > > problem > > > > > (or not) is: > > > > > > > > > > use Astro::Coord::ECI; > > > > > Astro::Coord::ECI->new()->get( 'sun' );

> > > > > > > > > > > > No test failures in Astro-satpass 0.102_01, versus 473 passes. I > > > > have > > > > published Astro-satpass 0.103, and closed RT 127488. I suspect > > > > this > > > > ticket can be closed also.

> > > > > > A possible action would be to adjust the minimum prereq version in > > > this distribution, so I'll leave the decision about closing this > > > ticket to the author. > > > > > > But I can confirm that the newer version indeed fixes the problem.

> > > >

> > I still see failures if an older Astro::satpass is installed. > > Unfortunately it seems that setting a higher version for a recommended > module does not force a reinstall.

Maybe ANDK has an idea for this problem? Show quoted text

> > Maybe the only solution is to also check the version in the test > script, and skip if it's the problematic version.

Thu Jan 31 16:34:12 2019 ANDK [...] cpan.org - Correspondence added

On 20190129T203631Z, SREZIC wrote: Show quoted text

> Maybe ANDK has an idea for this problem?

I usually say, if it can make a test fail, it is a dependency, not a recommendation. Looking closer, the moon.t test depends on a certain version of X but the package in general, usually, does not. So this test is not representative for the general use case. So I would recommend (sic!), to write the test in a way that makes it pass when the user has a wrong version of X installed. Generating a warning both during the test and during any execution that generates garbage due to a broken version of X would be the extra amount of politeness to the end user that could resolve any misunderstanding between the end user and the developer. Letting the test fail is a double bind: a recommendation with a separate bill on the backside for users who do not follow the recommendation, at least for some of them. But such extra warnings also come at a price and will probably never be removed even when after a few years no old versions of X will be around anymore. As the author I would probably be enough annoyed by FAIL reports to go that extra mile.