Skip Menu |
 

This queue is for tickets about the File-ShareDir CPAN distribution.

Report information
The Basics
Id: 86478
Status: stalled
Priority: 0/
Queue: File-ShareDir

People
Owner: Nobody in particular
Requestors: KENTNL [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: (no value)
Broken in: 1.03
Fixed in: (no value)



Subject: Suggestion: Drop support for legacy File::ShareDir's
The way FileShareDir works is that in the event a user accidentally fails to provide some kind of "File::ShareDir::Install" instruction to Makefile.PL/Build.PL, their installed dist will still falsely believe that the dist has a "working" share dir.

Its just confusingly empty, albeit for the packlist.

For instance, if you take a distribution known not to have any files installed in the sharedir,  File::ShareDir still returns a "valid" sharedir.

perl -MFile::ShareDir=dist_dir -wE 'say dist_dir(q{Package-Stash})'
/home/kent/perl5/perlbrew/perls/perl-5.18.0/lib/site_perl/5.18.0/x86_64-linux/auto/Package/Stash


perl -MPath::Tiny -MFile::ShareDir=dist_dir -wE 'say for path(dist_dir(q{Package-Stash}))->children'
/home/kent/perl5/perlbrew/perls/perl-5.18.0/lib/site_perl/5.18.0/x86_64-linux/auto/Package/Stash/.packlist
/home/kent/perl5/perlbrew/perls/perl-5.18.0/lib/site_perl/5.18.0/x86_64-linux/auto/Package/Stash/XS

Eliminating the _dist_dir_old mechansim would mean that instead of Dists being confused about where their files are, would instead be told "there is no dist dir" , which makes it more obvious that they've broken their install chain.
Download (untitled) / with headers
text/plain 102b
Yes please, it's high time to drop support for the old system. Let's do this in a major version bump.
Download (untitled) / with headers
text/plain 311b
Ether, sometimes a small smell of "kill someone before stepping to work" comes around when I read such follow-ups. You got File::ShareDir::Install and you know about the trouble around FHS and the "right" way and that we wanted to discuss all of that in Berlin. So what in hell brought you to push that button?

On 2015-02-23 00:55:22, REHSACK wrote:
> Ether, sometimes a small smell of "kill someone before stepping to
> work" comes around when I read such follow-ups. You got
> File::ShareDir::Install and you know about the trouble around FHS and
> the "right" way and that we wanted to discuss all of that in Berlin.
>
> So what in hell brought you to push that button?

I'm not sure what ETHER said that garnered this (apparently emotional) response.

However, I'm really looking forward to being filled in on what the specifics are of the FHS problem, because A) I care much about this problem and B) I won't be at Berlin.

If, for instance, that the proposed "right" FHS layout will be identical to the "old" layout, then I will have to register some opposition to *that* solution, because *that* solution
is presently the target for .packlist files, and every dist has such a dir, and as such
there's no way to distinguish between an "existing" and "non existing" dist dir.

But I'm not saying that *is* what you're considering, I'm just sitting here patiently waiting to hear the proposal(s) so I can participate in appropriately giving useful feedback.

But I will be very disappointed if I can't take part in the relevant discussion due to simply not being able to attend berlin.
 

Download (untitled) / with headers
text/plain 104b
I would love to see a patch with reasonable test and POD update for this ticket, if it's still relevant.


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.