Skip Menu |
 

This queue is for tickets about the Module-Build CPAN distribution.

Report information
The Basics
Id: 110882
Status: resolved
Priority: 0/
Queue: Module-Build

People
Owner: Nobody in particular
Requestors: ribasushi [...] leporine.io
Cc: john.peacock [...] havurah-software.org
AdminCc:

Bug Information
Severity: Critical
Broken in: (no value)
Fixed in: (no value)



CC: john.peacock [...] havurah-software.org
Subject: Test failure when version@0.9912_01 is in @INC
Download (untitled) / with headers
text/plain 519b
t/properties/release_status.t .. 1/? # Failed test 'dotted alpha version has release_status 'testing'' # at t/properties/release_status.t line 78. # got: 'stable' # expected: 'testing' t/properties/release_status.t .. 11/? # Looks like you failed 1 test of 21. ... Files=53, Tests=1132, 51 wallclock secs ( 0.54 usr 0.08 sys + 42.61 cusr 4.93 csys = 48.16 CPU) Result: FAIL Failed 1/53 test programs. 1/1132 subtests failed. Downgrading to the last stable version.pm 0.9912 makes tests pass again.
Download (untitled) / with headers
text/plain 827b
On Sun Jan 03 17:08:04 2016, RIBASUSHI wrote: Show quoted text
> t/properties/release_status.t .. 1/? > # Failed test 'dotted alpha version has release_status 'testing'' > # at t/properties/release_status.t line 78. > # got: 'stable' > # expected: 'testing' > t/properties/release_status.t .. 11/? # Looks like you failed 1 test > of 21.
The only reason why an alpha release of version.pm should be installed is to test modules that use it (like M::B), where the appropriate fix would be to ignore that error for the duration of your tests. Once 0.9913 is released (i.e. stable), this aberration will vanish. I'm also not convinced that M::B has any business throwing an error (rather than a warning) when an alpha release of version is installed, even though M::B depends closely on version.pm for core functionality. John
Subject: Re: [rt.cpan.org #110882] Test failure when version@0.9912_01 is in @INC
Date: Mon, 4 Jan 2016 00:26:24 +0100
To: bug-Module-Build [...] rt.cpan.org
From: Leon Timmermans <leont [...] cpan.org>
Download (untitled) / with headers
text/plain 1.2k
On Sun, Jan 3, 2016 at 11:08 PM, Peter Rabbitson via RT < bug-Module-Build@rt.cpan.org> wrote: Show quoted text
> Sun Jan 03 17:08:04 2016: Request 110882 was acted upon. > Transaction: Ticket created by RIBASUSHI > Queue: Module-Build > Subject: Test failure when version@0.9912_01 is in @INC > Broken in: (no value) > Severity: Critical > Owner: Nobody > Requestors: ribasushi@cpan.org > Status: new > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=110882 > > > > t/properties/release_status.t .. 1/? > # Failed test 'dotted alpha version has release_status 'testing'' > # at t/properties/release_status.t line 78. > # got: 'stable' > # expected: 'testing' > t/properties/release_status.t .. 11/? # Looks like you failed 1 test of 21. > ... > Files=53, Tests=1132, 51 wallclock secs ( 0.54 usr 0.08 sys + 42.61 cusr > 4.93 csys = 48.16 CPU) > Result: FAIL > Failed 1/53 test programs. 1/1132 subtests failed. > > > Downgrading to the last stable version.pm 0.9912 makes tests pass again. >
This seems a fairly direct consequence of the Lyon decision that «stringify… should include underscores only if the initializing value was a non-vstring string.». I'm not sure how to best resolve this, but I've reported this upstream for advice. Leon
CC: bug-Module-Build [...] rt.cpan.org
Subject: Re: [rt.cpan.org #110882] Test failure when version@0.9912_01 is in @INC
Date: Mon, 04 Jan 2016 09:21:37 +0100
To: john.peacock [...] havurah-software.org
From: Peter Rabbitson <ribasushi [...] cpan.org>
Download (untitled) / with headers
text/plain 648b
On 01/03/2016 11:49 PM, John Peacock via RT wrote: Show quoted text
> The only reason why an alpha release of version.pm should be installed is to test modules that use it (like M::B), where the appropriate fix would be to ignore that error for the duration of your tests.
John, in http://www.nntp.perl.org/group/perl.perl5.porters/2016/01/msg233584.html you say Show quoted text
> I will probably release 0.9912_01 to CPAN in the next day or two ... I > will broadcast that from the roof tops (and IRC channels) to get some > eyes on it ASAP.
This bugreport is me "getting some eyes on it" via a proper depchain smoke, similar to what ANDK does on a larger scale. Cheers
Subject: Re: [rt.cpan.org #110882] Test failure when version@0.9912_01 is in @INC
Date: Mon, 04 Jan 2016 09:23:24 +0100
To: bug-Module-Build [...] rt.cpan.org
From: Peter Rabbitson <ribasushi [...] cpan.org>
Download (untitled) / with headers
text/plain 216b
On 01/04/2016 12:27 AM, Leon Timmermans via RT wrote: Show quoted text
> I'm not sure how to best resolve this, but I've reported this upstream > for advice.
Could you link the related discussion so others can follow it? Cheers!
Download (untitled) / with headers
text/plain 145b
Show quoted text
> > Could you link the related discussion so others can follow it? >
There it is: https://gist.github.com/dagolden/9559280#gistcomment-1661861
RT-Send-CC: LEONT [...] cpan.org, john.peacock [...] havurah-software.org
On 2016-01-03 17:49:44, JPEACOCK wrote: Show quoted text
> On Sun Jan 03 17:08:04 2016, RIBASUSHI wrote:
> > t/properties/release_status.t .. 1/? > > # Failed test 'dotted alpha version has release_status 'testing'' > > # at t/properties/release_status.t line 78. > > # got: 'stable' > > # expected: 'testing' > > t/properties/release_status.t .. 11/? # Looks like you failed 1 test > > of 21.
> > The only reason why an alpha release of version.pm should be installed > is to test modules that use it (like M::B), where the appropriate fix > would be to ignore that error for the duration of your tests. Once > 0.9913 is released (i.e. stable), this aberration will vanish.
I have to disagree. Devel releases should work as good as stable ones, for all modules, period. And where you put the border line for "using it", for such a core module like version.pm? metacpan.org lists 1249 CPAN distributions directly depending on version.pm, and deps.cpantesters.org lists > 40000 distributions indirectly depending on it (probably all CPAN distributions).
Download (untitled) / with headers
text/plain 1010b
On Mon Jan 04 14:31:50 2016, SREZIC wrote: Show quoted text
> I have to disagree. Devel releases should work as good as stable ones, > for all modules, period.
As an aside, that is a very short-sighted view, IMNSO. Devel releases might include massive refactoring and could be reasonably expected to break at least some modules. Devel releases might just be a preview of new functionality and should not disturb existing consumers. You shouldn't assume anything without checking the README or POD for the module in question. However, I misspoke in my original comment. I though M::B was complaining about version.pm itself being an alpha release, not that the behavior of stringified dotted decimal versions was acting differently. I would have expected this alpha release of version.pm to not break anything, since it was a fairly narrow change affecting one particular edge case that most people will never stumble over. I was wrong, but pending the groupmind's discussion, I don't know which way to "fix" things.
Subject: Re: [rt.cpan.org #110882] Test failure when version@0.9912_01 is in @INC
Date: Mon, 4 Jan 2016 20:49:42 -0600
To: bug-Module-Build [...] rt.cpan.org
From: John Peacock <john.peacock [...] havurah-software.org>
Download (untitled) / with headers
text/plain 325b
On 01/04/2016 02:22 AM, Peter Rabbitson via RT wrote: Show quoted text
> This bugreport is me "getting some eyes on it" via a proper depchain > smoke, similar to what ANDK does on a larger scale.
Right, but I think that test needs to be turned into a TODO until such time as the decision is made what to do about that jagged edge... John
RT-Send-CC: LEONT [...] cpan.org, LEONT [...] cpan.org, john.peacock [...] havurah-software.org
Download (untitled) / with headers
text/plain 1.4k
On 2016-01-04 21:45:17, JPEACOCK wrote: Show quoted text
> On Mon Jan 04 14:31:50 2016, SREZIC wrote:
> > I have to disagree. Devel releases should work as good as stable > > ones, > > for all modules, period.
> > As an aside, that is a very short-sighted view, IMNSO. Devel releases > might include massive refactoring and could be reasonably expected to > break at least some modules. Devel releases might just be a preview > of new functionality and should not disturb existing consumers. You > shouldn't assume anything without checking the README or POD for the > module in question.
If there's a known problem happening only with devel versions (and this every time!) then this problem should be fixed. We want to be able to do statistical analysis of cpantesters reports, and such fails ruin these attempts. This was my main objection to your statement. For unexpected problems caused by devel releases... yes, devel versions may have bugs, and it's better if they are caught before doing the stable release. Still these problems could add additional "noise" to the cpantesters data, making it more difficult to analyze reports (e.g. for regression analysis for bleadperl). Yes, the cpantesters system is lacking something here, maybe it should just be possible to put a mark on smokers which are using devel module versions, so we can filter reports if needed for a specific analysis. Until we're there it surely helps if devel versions are not that experimental, and do not cause unnecessary failures.
Download (untitled) / with headers
text/plain 470b
On Mon Jan 04 00:26:59 2016, LEONT wrote: Show quoted text
> > This seems a fairly direct consequence of the Lyon decision that > «stringify… should include underscores only if the initializing value > was a > non-vstring string.». I'm not sure how to best resolve this, but I've > reported this upstream for advice.
Any progress on this? Both https://gist.github.com/dagolden/9559280 and http://www.nntp.perl.org/group/perl.perl5.porters/2016/01/msg233584.html have gone silent.
Subject: Re: [rt.cpan.org #110882] Test failure when version@0.9912_01 is in @INC
Date: Wed, 20 Jan 2016 10:52:49 +0100
To: bug-Module-Build [...] rt.cpan.org
From: Leon Timmermans <leont [...] cpan.org>
Download (untitled) / with headers
text/plain 795b
On Sun, Jan 10, 2016 at 4:23 PM, Peter Rabbitson via RT < bug-Module-Build@rt.cpan.org> wrote: Show quoted text
> Queue: Module-Build > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=110882 > > > On Mon Jan 04 00:26:59 2016, LEONT wrote:
> > > > This seems a fairly direct consequence of the Lyon decision that > > «stringify… should include underscores only if the initializing value > > was a > > non-vstring string.». I'm not sure how to best resolve this, but I've > > reported this upstream for advice.
> > Any progress on this? Both https://gist.github.com/dagolden/9559280 and > http://www.nntp.perl.org/group/perl.perl5.porters/2016/01/msg233584.html > have gone silent. >
My apologies for the delay, I've just released a version of Module::Build without that specific test. Leon
Subject: Re: [rt.cpan.org #110882] Test failure when version@0.9912_01 is in @INC
Date: Wed, 20 Jan 2016 10:28:14 -0600
To: bug-Module-Build [...] rt.cpan.org, ribasushi [...] cpan.org
From: John Peacock <john.peacock [...] havurah-software.org>
Download (untitled) / with headers
text/plain 256b
On 01/20/2016 03:53 AM, Leon Timmermans via RT wrote: Show quoted text
> My apologies for the delay, I've just released a version of Module::Build > without that specific test.
So I am going to consider this no longer a blocker to version.pm 0.9913 being released. John


This service is sponsored and maintained by Best Practical Solutions and runs on Perl.org infrastructure.

Please report any issues with rt.cpan.org to rt-cpan-admin@bestpractical.com.