develooper Front page | perl.perl5.porters | Postings from July 2020

Re: A thought about 5-7 compatibility

Thread Previous | Thread Next
From:
Paul "LeoNerd" Evans
Date:
July 2, 2020 22:00
Subject:
Re: A thought about 5-7 compatibility
Message ID:
20200702225955.2d83dc64@shy.leonerd.org.uk
On Thu, 2 Jul 2020 23:17:30 +0300
Veesh Goldman <rabbiveesh@gmail.com> wrote:

> I'm curious what everyone thinks about this idea to solve the CPAN
> backcompat issue.
> Perhaps at module install time, the installer attempts to run the
> tests for the module being installed under the assumption that the
> module is perl7. If that fails, then on the system mark that the
> module is perl5, and store some metadata (somewhere) that the module
> must be in p5 compatibility mode.
> 
> Of course, this would require changing the installers (which probably
> need updating anyways) and the addition of some metadata store
> (unless there is one already).

I think there would be far too many false positives and false negatives
here to offer any meaningful signal.

Many dists (maaaaany dists) don't even have any tests at all, so I
wouldn't take the lack of failure under perl7 to mean they're
perl7-safe. Also many dists will have spurious failures all over the
place - e.g. right now I have a bunch of fails of Object::Pad on perls
5.16 to 5.20 due to some uninitialised pointer I've yet to properly
track down. I'd hate to have such failures randomly decide which install
path it gets installed as.

You're basically flipping a coin at install time in order to pick one
of two paths, which will hugely matter. It will make inspection and
debugging of these things much harder.

In summary: I'm against it.

-- 
Paul "LeoNerd" Evans

leonerd@leonerd.org.uk      |  https://metacpan.org/author/PEVANS
http://www.leonerd.org.uk/  |  https://www.tindie.com/stores/leonerd/

Thread Previous | Thread Next


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About