Bug #113948 for Tree-Simple: Test::* should be a testing dependency, not a runtime requirement

Sun Apr 24 04:51:05 2016 KENTNL [...] cpan.org - Ticket created

Subject:

Test::* should be a testing dependency, not a runtime requirement

https://metacpan.org/source/RSAVAGE/Tree-Simple-1.26/Makefile.PL#L33-34

https://metacpan.org/source/RSAVAGE/Tree-Simple-1.26/META.json#L36-37

https://metacpan.org/source/RSAVAGE/Tree-Simple-1.26/META.yml#L22-23

Sun Apr 24 19:14:45 2016 ron [...] savage.net.au - Correspondence added

Subject:	Re: [rt.cpan.org #113948] Test::* should be a testing dependency, not a runtime requirement
Date:	Mon, 25 Apr 2016 09:14:25 +1000
To:	bug-Tree-Simple [...] rt.cpan.org
From:	Ron Savage <ron [...] savage.net.au>

Hi Kent On 24/04/16 18:51, Kent Fredric via RT wrote: Show quoted text

> Sun Apr 24 04:51:05 2016: Request 113948 was acted upon. > Transaction: Ticket created by KENTNL > Queue: Tree-Simple > Subject: Test::* should be a testing dependency, not a runtime requirement > Broken in: 1.26 > Severity: (no value) > Owner: Nobody > Requestors: KENTNL@cpan.org > Status: new > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=113948 > > > > https://metacpan.org/source/RSAVAGE/Tree-Simple-1.26/Makefile.PL#L33-34 > > https://metacpan.org/source/RSAVAGE/Tree-Simple-1.26/META.json#L36-37 > > https://metacpan.org/source/RSAVAGE/Tree-Simple-1.26/META.yml#L22-23

Thanx for this report. I got a shock when I read it because I just knew how few of the modules I wrote or adopted use TEST_REQUIRES. E.g. Out of 8 Tree::* modules only 1 does - Tree::DAG_Node. Of course I should have been fixing this as I worked on them, the same way I've been wiping out Build.PL per release. Also fiendishly clever of you to send this on an Australian national holiday so I have the free time to start fixing 129 modules :-). -- Ron Savage - savage.net.au

Sun Apr 24 19:14:45 2016 The RT System itself - Status changed from 'new' to 'open'

Mon Apr 25 03:33:44 2016 RSAVAGE [...] cpan.org - Correspondence added 30 min

V 1.27 is on CPAN.

Mon Apr 25 03:33:45 2016 RSAVAGE [...] cpan.org - Status changed from 'open' to 'resolved'

Mon Apr 25 03:33:45 2016 RSAVAGE [...] cpan.org - Fixed in 1.27 added

Mon Apr 25 23:05:06 2016 RSAVAGE [...] cpan.org - Correspondence added

Hi Kent I've been informed that the META.json file generated does not include the repository info from Makefile.PL. I've asked for help on the CPAN-testers-discuss mailing list, so I expect I'll be releasing a new version shortly.

Mon Apr 25 23:57:11 2016 KENTNL [...] cpan.org - Correspondence added

On 2016-04-26 15:05:06, RSAVAGE wrote:
> Hi Kent
>
> I've been informed that the META.json file generated does not include
> the repository info from Makefile.PL. I've asked for help on the CPAN-
> testers-discuss mailing list, so I expect I'll be releasing a new
> version shortly.

If I had to guess, I'd say it might be related to no code in your Makefile.PL enforcing META-2.0, so its probably assuming META-1.0 semantics, losing that data, and then up-converting the lost data to META-2.0 for META.json.

But I'm not sure, try it and see.

Tue Apr 26 00:01:52 2016 KENTNL [...] cpan.org - Correspondence added

On 2016-04-26 15:57:11, KENTNL wrote:
>
> If I had to guess, I'd say it might be related to no code in your
> Makefile.PL
> enforcing META-2.0, so its probably assuming META-1.0 semantics,
> losing that
> data, and then up-converting the lost data to META-2.0 for META.json.
>
> But I'm not sure, try it and see.

This for instance appears to work ... somewhat. The extra keys being lost however I can't explain.

https://metacpan.org/source/RSAVAGE/GraphViz2-Marpa-PathUtils-2.00/Makefile.PL#L120-123

https://metacpan.org/source/RSAVAGE/GraphViz2-Marpa-PathUtils-2.00/META.json#L75

Tue Apr 26 00:05:30 2016 KENTNL [...] cpan.org - Correspondence added

Looking at your source repo, I think you have a workflow problem.

META.* should not be in the repository, but they are.

MYMETA.* should CERTAINLY not be in the repository, but they are.

Those files should be generated during `make dist`, and never exist in your work tree, under a standard EUMM workflow.

Tue Apr 26 01:06:51 2016 ron [...] savage.net.au - Correspondence added

Subject:	Re: [rt.cpan.org #113948] Test::* should be a testing dependency, not a runtime requirement
Date:	Tue, 26 Apr 2016 15:06:31 +1000
To:	bug-Tree-Simple [...] rt.cpan.org
From:	Ron Savage <ron [...] savage.net.au>

Hi Kent On 26/04/16 14:05, Kent Fredric via RT wrote: Show quoted text

> Queue: Tree-Simple > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=113948 > > > Looking at your source repo, I think you have a workflow problem.

Agreed. Show quoted text

> META.* should not be in the repository, but they are. > > MYMETA.* should CERTAINLY not be in the repository, but they are. > > Those files should be generated during `make dist`, and never exist in your > work tree, under a standard EUMM workflow.

You pull worked perfectly. I'll make the other repos look-alikes. $many x $thanx; -- Ron Savage - savage.net.au