Skip Menu |

This queue is for tickets about the CPANPLUS CPAN distribution.

Report information
The Basics
Id: 27911
Status: rejected
Priority: 0/
Queue: CPANPLUS

People
Owner: Nobody in particular
Requestors: doughera [...] lafayette.edu
Cc:
AdminCc:

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



CC: Perl Porters <perl5-porters [...] perl.org>
Subject: CPANPLUS tests incorrectly write to my real perl directory
Date: Mon, 2 Jul 2007 15:48:54 -0400 (EDT)
To: bug-cpanplus [...] rt.cpan.org
From: Andy Dougherty <doughera [...] lafayette.edu>
While testing CPANPLUS on my bleadperl test directory, I ran into the following problem. I built and installed bleadperl into a dedicated directory. I then ran cpanp5.9.5 and told it to t CPANPLUS it dutifully fetched CPANPLUS, but during the testing phase of t/20_CPLANPLUS-Dist-MM I saw the following error message: # Sorry, installing into your real perl dir, rather than our test # area since ExtUtils::Installed does not probe for .packlists in # other dirs than those in %Config. See bug #6871 on rt.cpan.org # for details This is wrong. The test should be skipped. Or I should be prompted to agree to polluting my 'real perl dir'. Now I have to go in and manually clean up whatever stuff it put there. I shouldn't have to do this. Testing is not installing. -- Andy Dougherty doughera@lafayette.edu
On Jul 2, 2007, at 9:48 PM, Andy Dougherty wrote: Show quoted text
> While testing CPANPLUS on my bleadperl test directory, I ran into the > following problem. > > I built and installed bleadperl into a dedicated directory. I then > ran > cpanp5.9.5 and told it to > t CPANPLUS > it dutifully fetched CPANPLUS, but during the testing phase of > t/20_CPLANPLUS-Dist-MM > > I saw the following error message: > > # Sorry, installing into your real perl dir, rather than our test > # area since ExtUtils::Installed does not probe for .packlists in > # other dirs than those in %Config. See bug #6871 on rt.cpan.org > # for details > > This is wrong. The test should be skipped. Or I should be > prompted to > agree to polluting my 'real perl dir'. Now I have to go in and > manually > clean up whatever stuff it put there. I shouldn't have to do this. > Testing is not installing.
It installs, and removes it and only when there's sufficient permissions, so there's nothing to actually to worry about. If you read down the test diagnostics, you will see: # Sorry, installing into your real perl dir, rather than our test # area since ExtUtils::Installed does not probe for .packlists in # other dirs than those in %Config. See bug #6871 on rt.cpan.org # for details ok 19 - Installing module [...] ok 37 - Uninstalling module It's always been deemend a worthwhile test, even though the recorded bug in EU::Installed. If this is actually now considered a big problem, I suggest we put real effort in fixing the reported bug in EU::Installed, rather than throwing away important feature tests. For the record: $ corelist ExtUtils::Installed ExtUtils::Installed was first released with perl 5.005 And the bug has been open for 2 years now. -- Jos Boumans
Subject: Re: [rt.cpan.org #27911] CPANPLUS tests incorrectly write to my real perl directory
Date: Tue, 3 Jul 2007 09:53:02 -0400 (EDT)
To: via RT <bug-CPANPLUS [...] rt.cpan.org>
From: Andy Dougherty <doughera [...] lafayette.edu>
On Tue, 3 Jul 2007, via RT wrote: Show quoted text
> > <URL: http://rt.cpan.org/Ticket/Display.html?id=27911 > > > On Jul 2, 2007, at 9:48 PM, Andy Dougherty wrote: >
> > While testing CPANPLUS on my bleadperl test directory, I ran into the > > following problem. > > > > I built and installed bleadperl into a dedicated directory. I then > > ran > > cpanp5.9.5 and told it to > > t CPANPLUS > > it dutifully fetched CPANPLUS, but during the testing phase of > > t/20_CPLANPLUS-Dist-MM > > > > I saw the following error message: > > > > # Sorry, installing into your real perl dir, rather than our test > > # area since ExtUtils::Installed does not probe for .packlists in > > # other dirs than those in %Config. See bug #6871 on rt.cpan.org > > # for details > > > > This is wrong. The test should be skipped. Or I should be > > prompted to > > agree to polluting my 'real perl dir'. Now I have to go in and > > manually > > clean up whatever stuff it put there. I shouldn't have to do this. > > Testing is not installing.
> > It installs, and removes it and only when there's sufficient > permissions, > so there's nothing to actually to worry about.
Yes there is. First, from a fundamental point of view, I think that testing is not installation. Touching my installation directory without my consent really strikes me as wrong. Show quoted text
> If you read down the test diagnostics, you will see: > # Sorry, installing into your real perl dir, rather than our test > # area since ExtUtils::Installed does not probe for .packlists in > # other dirs than those in %Config. See bug #6871 on rt.cpan.org > # for details > ok 19 - Installing module > [...] > ok 37 - Uninstalling module
Yes, I saw that. But it doesn't actually completely uninstall the module. Before running the test, I made a fresh copy of my perl directory. After testing, I ran a recursive diff. I found that I still have the following entities that I didn't have before: perl/lib/5.9.5/i686-linux/perllocal.pod perl/lib/site_perl/5.9.5/Foo/ perl/lib/site_perl/5.9.5/i686-linux/auto/ The perllocal.pod one contains fake information about files that aren't installed anymore. Also, the timestamps on my installation directories are now changed. [This is all with CPANPLUS 0.80.] In short, this test changed my installation directories when I specificially had not asked for an installation. Show quoted text
> It's always been deemend a worthwhile test, even though the recorded bug > in EU::Installed. If this is actually now considered a big problem, I > suggest we put real effort in fixing the reported bug in EU::Installed, > rather than throwing away important feature tests.
I had already read the bug report referenced in the message, and I saw that it's a workaround for a problem in EU::Installed. I just disagree with your conclusion. I don't know how others feel. I may be a minority of one. -- Andy Dougherty doughera@lafayette.edu