Skip Menu |

This queue is for tickets about the Moose CPAN distribution.

Report information
The Basics
Id: 66211
Status: resolved
Priority: 0/
Queue: Moose
People
Owner: Nobody in particular
Requestors: ANDK [...] cpan.org
Cc:
AdminCc:
Bug Information
Severity: Normal
Broken in:
  • 1.9900-TRIAL
  • 1.9901-TRIAL
  • 1.9902-TRIAL
Fixed in: (no value)

History Show all quoted text
  Sat Feb 26 16:35:18 2011 ANDK [...] cpan.org - Ticket created
Subject: The 1.990x-TRIAL releases on CPAN should be deleted
I apologize that I do not closely follow the Moose development and have no idea what the purpose of the three trial releases is. But what I see is that they do not follow the established conventions on CPAN. All tools around CPAN are based on the idea that stable releases and dev releases follow a monotonically increasing versioning. Between two stable releases one can expect some dev releases. They have an underscore somewhere in the version or recently the idea was propagated that -TRIAL be part of the tarball name. Between 0.62 and 0,63 I see 0.62_01 and 0.61_02. Between 0.55 and 0.56 there were 4 dev releases. And then I see stable releases 1.21 (November), 1,22, 1,23, 1.24 (all February) and the dev releases in between? Instead 3 dev releases somewhere between 1.21 and 1.24 I find 1.990{0,1,2} in January. All tools that pay attention to dev releases are now expecting something Show quoted text
> 1.9902 as the next stable version but it doesn't arrive.
This was probably an oversight? I'd wish that Moose' next stable release would be > 1.9902 or if this suggestion doesn't find friends, that 1.990x be deleted from CPAN to prevent further damage. Damage is when hundreds of cpantester reports are generated testing 1.9902 and nobody is really interested in them. Please let me know when I grossly misunderstand what's going on here. Thanks,
  Mon Mar 14 20:32:04 2011 DROLSKY [...] cpan.org - Correspondence added
On Sat Feb 26 16:35:18 2011, ANDK wrote: Show quoted text
> I apologize that I do not closely follow the Moose development and have > no idea what the purpose of the three trial releases is. But what I see > is that they do not follow the established conventions on CPAN. > > All tools around CPAN are based on the idea that stable releases and dev > releases follow a monotonically increasing versioning. Between two > stable releases one can expect some dev releases. They have an > underscore somewhere in the version or recently the idea was propagated > that -TRIAL be part of the tarball name. > > Between 0.62 and 0,63 I see 0.62_01 and 0.61_02. Between 0.55 and 0.56 > there were 4 dev releases. > > And then I see stable releases 1.21 (November), 1,22, 1,23, 1.24 (all > February) and the dev releases in between? Instead 3 dev releases > somewhere between 1.21 and 1.24 I find 1.990{0,1,2} in January. > > All tools that pay attention to dev releases are now expecting something
> > 1.9902 as the next stable version but it doesn't arrive.
> > This was probably an oversight? I'd wish that Moose' next stable release > would be > 1.9902 or if this suggestion doesn't find friends, that > 1.990x be deleted from CPAN to prevent further damage.
The releases were released intentionally with the version numbers they have. Show quoted text
> Damage is when hundreds of cpantester reports are generated testing > 1.9902 and nobody is really interested in them.
Shouldn't tester boxen ignore TRIAL dists? Isn't that the real underlying problem? I don't think we're going to delete the TRIAL dists on CPAN, though we do hope to release an actual 2.0 soon(ish). -dave
  Mon Mar 14 20:32:05 2011 The RT System itself - Status changed from 'new' to 'open'
  Tue Mar 15 04:06:40 2011 andreas.koenig.7os6VVqR [...] franz.ak.mind.de - Correspondence added
CC: ANDK [...] cpan.org
Subject: Re: [rt.cpan.org #66211] The 1.990x-TRIAL releases on CPAN should be deleted
Date: Tue, 15 Mar 2011 09:06:18 +0100
To: bug-Moose [...] rt.cpan.org
From: andreas.koenig.7os6VVqR [...] franz.ak.mind.de (Andreas J. Koenig)
Show quoted text
>>>>> On Mon, 14 Mar 2011 20:32:05 -0400, "Dave Rolsky via RT" <bug-Moose@rt.cpan.org> said:
Show quoted text
>> Damage is when hundreds of cpantester reports are generated testing >> 1.9902 and nobody is really interested in them.
Show quoted text
> Shouldn't tester boxen ignore TRIAL dists? Isn't that the real > underlying problem?
If you are in the camp that thinks so, please read http://wiki.cpantesters.org/wiki/CPANAuthorNotes and add a line such as exit 0 if $ENV{AUTOMATED_TESTING}; to your Makefile.PL I think it is a very unexpected stance that somebody makes a strictly-not-for-testing trial release. What is it good for? Who's the target audience? Well, given that there is a way to opt out, you do not have to answer this question but it's still helpful if you explain your reasoning about it so that the rest of the world knows what you're intending with the releases. Show quoted text
> I don't think we're going to delete the TRIAL dists on CPAN, though we > do hope to release an actual 2.0 soon(ish).
1.9904 is actually much better than 1.9902. I still don't know should it be tested at all and by whom. -- andreas
  Tue Mar 15 10:52:38 2011 autarch [...] urth.org - Correspondence added
Subject: Re: [rt.cpan.org #66211] The 1.990x-TRIAL releases on CPAN should be deleted
Date: Tue, 15 Mar 2011 09:52:25 -0500 (CDT)
To: "(Andreas J. Koenig) via RT" <bug-Moose [...] rt.cpan.org>
From: Dave Rolsky <autarch [...] urth.org>
On Tue, 15 Mar 2011, (Andreas J. Koenig) via RT wrote: Show quoted text
> I think it is a very unexpected stance that somebody makes a > strictly-not-for-testing trial release. What is it good for? Who's the > target audience? Well, given that there is a way to opt out, you do not > have to answer this question but it's still helpful if you explain your > reasoning about it so that the rest of the world knows what you're > intending with the releases.
Err, that wasn't really clear. I know they should test the dist, but I don't understand why they would _install_ the dist when testing things that depend on Moose. Show quoted text
> > I don't think we're going to delete the TRIAL dists on CPAN, though we > > do hope to release an actual 2.0 soon(ish).
> > 1.9904 is actually much better than 1.9902. I still don't know should it > be tested at all and by whom.
Better how? -dave /*============================================================ http://VegGuide.org http://blog.urth.org Your guide to all that's veg House Absolute(ly Pointless) ============================================================*/
  Wed Mar 16 02:41:00 2011 andreas.koenig.7os6VVqR [...] franz.ak.mind.de - Correspondence added
CC: ANDK [...] cpan.org
Subject: Re: [rt.cpan.org #66211] The 1.990x-TRIAL releases on CPAN should be deleted
Date: Wed, 16 Mar 2011 07:40:32 +0100
To: bug-Moose [...] rt.cpan.org
From: andreas.koenig.7os6VVqR [...] franz.ak.mind.de (Andreas J. Koenig)
Show quoted text
>>>>> On Tue, 15 Mar 2011 10:52:39 -0400, "autarch@urth.org via RT" <bug-Moose@rt.cpan.org> said:
Show quoted text
> Err, that wasn't really clear. I know they should test the dist, but I > don't understand why they would _install_ the dist when testing things > that depend on Moose.
Testing in the broader sense is every activity that is aimed at finding bugs. Whereas only running a test suite is a pretty narrow view of searching for excellence. By testing dependents one usually finds more bugs because most test suites are either deliberately incomplete or simply not perfect. Show quoted text
>> > I don't think we're going to delete the TRIAL dists on CPAN, though we >> > do hope to release an actual 2.0 soon(ish).
>> >> 1.9904 is actually much better than 1.9902. I still don't know should it >> be tested at all and by whom.
Show quoted text
> Better how?
I expressed it completely wrong. Rephrasing: My impression is that less third party Moose dependents show up as broken. The reason is not clear at the moment. It can either be that fewer people installed 1.9904 for testing dependents or that 1.9904 is really better. To answer that question would be a huge statistical effort that I can't provide at the moment. -- andreas
  Wed Mar 16 17:40:24 2011 autarch [...] urth.org - Correspondence added
Subject: Re: [rt.cpan.org #66211] The 1.990x-TRIAL releases on CPAN should be deleted
Date: Wed, 16 Mar 2011 16:40:14 -0500 (CDT)
To: "(Andreas J. Koenig) via RT" <bug-Moose [...] rt.cpan.org>
From: Dave Rolsky <autarch [...] urth.org>
On Wed, 16 Mar 2011, (Andreas J. Koenig) via RT wrote: Show quoted text
> Queue: Moose > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=66211 > >
>>>>>> On Tue, 15 Mar 2011 10:52:39 -0400, "autarch@urth.org via RT" <bug-Moose@rt.cpan.org> said:
>
> > Err, that wasn't really clear. I know they should test the dist, but I > > don't understand why they would _install_ the dist when testing things > > that depend on Moose.
> > Testing in the broader sense is every activity that is aimed at finding > bugs. Whereas only running a test suite is a pretty narrow view of > searching for excellence. By testing dependents one usually finds more > bugs because most test suites are either deliberately incomplete or > simply not perfect.
Sure, I suppose. In our case, we knew that Moose 2.0 broke a bunch of this stuff, so the releases are really for people to have something to _fix_ their code against. Show quoted text
> My impression is that less third party Moose dependents show up as > broken. The reason is not clear at the moment. It can either be that > fewer people installed 1.9904 for testing dependents or that 1.9904 is > really better. To answer that question would be a huge statistical > effort that I can't provide at the moment.
That's mostly because the relevant MooseX modules have been updated, I think. Regardless, I _still_ don't understand your original request that the TRIAL releases be deleted. -dave /*============================================================ http://VegGuide.org http://blog.urth.org Your guide to all that's veg House Absolute(ly Pointless) ============================================================*/
  Fri Mar 18 03:46:06 2011 andreas.koenig.7os6VVqR [...] franz.ak.mind.de - Correspondence added
CC: ANDK [...] cpan.org
Subject: Re: [rt.cpan.org #66211] The 1.990x-TRIAL releases on CPAN should be deleted
Date: Fri, 18 Mar 2011 08:45:45 +0100
To: bug-Moose [...] rt.cpan.org
From: andreas.koenig.7os6VVqR [...] franz.ak.mind.de (Andreas J. Koenig)
Show quoted text
>> Testing in the broader sense is every activity that is aimed at finding >> bugs. Whereas only running a test suite is a pretty narrow view of >> searching for excellence. By testing dependents one usually finds more >> bugs because most test suites are either deliberately incomplete or >> simply not perfect.
Show quoted text
> Sure, I suppose. In our case, we knew that Moose 2.0 broke a bunch of this > stuff, so the releases are really for people to have something to _fix_ > their code against.
Well, this statement changes the picture I have of the state of affairs. If you break dependents, you must have a strong interest in the results of tests combining current releases of dependents with dev releases of Moose-2.0-tobe. I would expect that you welcome or at least wouldn't reject reports about such breakages to aid in a smooth transition to Moose-2.0. But on IRC your comment on such tests was a succint (2010-12-18T16:18) < autarch> it doesn't make sense for us ;) Show quoted text
> Regardless, I _still_ don't understand your original request that the > TRIAL releases be deleted.
Because I was under the impression that there was less than zero interest in the FAIL reports sent out by cpantesters accounting which dependents would break with 1.9902. If you now declare that continuing testing 1.99* against dependent modules does make sense, then I can unblock my interrupted testing and there is no reason anymore to demand the removal. It's all just a matter of communicating the goals;) -- andreas
  Sun Mar 20 14:30:44 2011 autarch [...] urth.org - Correspondence added
Subject: Re: [rt.cpan.org #66211] The 1.990x-TRIAL releases on CPAN should be deleted
Date: Sun, 20 Mar 2011 13:30:34 -0500 (CDT)
To: "(Andreas J. Koenig) via RT" <bug-Moose [...] rt.cpan.org>
From: Dave Rolsky <autarch [...] urth.org>
On Fri, 18 Mar 2011, (Andreas J. Koenig) via RT wrote: Show quoted text
> Well, this statement changes the picture I have of the state of affairs. > If you break dependents, you must have a strong interest in the results > of tests combining current releases of dependents with dev releases of > Moose-2.0-tobe. I would expect that you welcome or at least wouldn't > reject reports about such breakages to aid in a smooth transition to > Moose-2.0.
Well, the reports are mostly just noise for us, I think. We have our own test harness for testing dependents that we use to screen for failures. Show quoted text
> If you now declare that continuing testing 1.99* against dependent > modules does make sense, then I can unblock my interrupted testing and > there is no reason anymore to demand the removal.
Either way. As I said, we have our own test harness for this, so additional failure reports for dependencies don't add much. Show quoted text
> It's all just a matter of communicating the goals;)
I've always thought the main purposes for a dev release are: * Test on other architecures then your own * Give downstream dependents something to code against before a stable release that breaks backwards compat The latter is also really important for DarkPAN, which is another reason to make dev releases, regardless of whether the test reports from that release are useful. -dave /*============================================================ http://VegGuide.org http://blog.urth.org Your guide to all that's veg House Absolute(ly Pointless) ============================================================*/
  Sun Mar 20 19:37:35 2011 DROLSKY [...] cpan.org - Status changed from 'open' to 'resolved'