Skip Menu |

This queue is for tickets about the Class-Method-Modifiers CPAN distribution.

Report information
The Basics
Id: 88641
Status: resolved
Priority: 0/
Queue: Class-Method-Modifiers

People
Owner: Nobody in particular
Requestors: ribasushi [...] leporine.io
Cc:
AdminCc:

Bug Information
Severity: Wishlist
Broken in: (no value)
Fixed in: 2.08



Subject: 5.6 compatibility promptly lost in dzil conversion
Please consider repackaging CMM using a more conventional dzil installer method (say EUMM), as opposed to Moduke::Build::Tiny. MBT requires 5.8 to function (due to over-relience on proper IO layers) Cheers
On 2013-09-12 21:58:49, RIBASUSHI wrote: Show quoted text
> Please consider repackaging CMM using a more conventional dzil > installer method (say EUMM), as opposed to Moduke::Build::Tiny. MBT > requires 5.8 to function (due to over-relience on proper IO layers) > > Cheers
My feeling is that we've gained more by avoiding EUMM and supporting leont's advances in this area, than we've lost by abandoning 5.6. I've heard of exactly *zero* users still on 5.6 that are upgrading modules from the CPAN. I'll leave this open however as it's not my unilateral decision.
it's an easy enough change, so I fixed CMM so it uses EUMM once more.
On Fri Sep 13 13:33:56 2013, ETHER wrote: Show quoted text
> it's an easy enough change, so I fixed CMM so it uses EUMM once more.
I was aware the change was a one-liner, which is why I didn't provide a patch. If there are no other immediate pending improvements, could you please cut a new PAUSE upload so that I can revert https://github.com/dbsrgits/dbix-class/commit/72cd4007 and as a result start smoking[1] the lvalue changes (though in fairness preliminary testing did not reveal issues) [1] I say smoking because the travis-ci setup of DBIC at times goes to far greater lengths than what is commonly done on cpantesters. Maybe I should just enable a reporter... ;)
2.06 is released using EUMM.
Just reopening this, so I get an email when the fixed version ships. If you need help with tarball testing on perl v.<old> - let me know.
2.08 is close but no cigar. It has two remaining issues with its deps, I am taking care of one of them, but the other is for you: you use the `if` module in all tests. It is not core until 5.6.2. Installing it on 5.6.1 goes through without a hitch, so it's a mere META ommission. Cheers! P.S. this is a minor issue, and can wait unti the release with mst's docpatches or whatnot P.P.S If you are wondering what the second issue is - MBT does not install on 5.6, but that's a patch-fest for another day, and not really your problem.
oops sorry didn't see your mail until just after I closed this.
On 2013-10-11 02:32:47, RIBASUSHI wrote: Show quoted text
> you use the `if` module in all tests. It is not core until 5.6.2. > Installing it on 5.6.1 goes through without a hitch, so it's a mere > META ommission.
Is your issue that it wasn't in core in 5.6.0, or...? because it is listed in META.* as a test requirement. Can you link to the MBT issue here?
On Fri Oct 11 11:55:01 2013, ETHER wrote: Show quoted text
> On 2013-10-11 02:32:47, RIBASUSHI wrote: >
> > you use the `if` module in all tests. It is not core until 5.6.2. > > Installing it on 5.6.1 goes through without a hitch, so it's a mere > > META ommission.
> > Is your issue that it wasn't in core in 5.6.0, or...? because it is > listed in META.* as a test requirement.
I am an idiot - I honestly missed that. With that the only issue remain MBT, so nothing else left on the 5.6 compat front as far as this dist is concerned. Show quoted text
> Can you link to the MBT issue here?
I haven't filed any - it's a general "depchain does not work on 5.6", and I am planning to look into that some time next week.