On 5 April 2017 at 10:35, James E Keenan <jkee...@pobox.com> wrote: > Module::Runtime was successfully installed all three times -- though that's > actually two times, because my first run used an earlier perl 5 blead > commit. Each of the three runs used a different, progressively newer > version of 'cpanm'. > > Task::Weaken did not succeed on my first run (20170330). It failed here: > > ##### > Can't locate inc/Module/Install.pm in @INC (you may need to install the > inc::Module::Install module) (@INC contains: > /home/jkeenan/testing/blead/lib/perl5/site_perl/5.25.11/x86_64-linux > /home/jkeenan/testing/blead/lib/perl5/site_perl/5.25.11 > /home/jkeenan/testing/blead/lib/perl5/5.25.11/x86_64-linux > /home/jkeenan/testing/blead/lib/perl5/5.25.11) at Makefile.PL line 42. > BEGIN failed--compilation aborted at Makefile.PL line 42. > -> N/A > -> FAIL Configure failed for Task-Weaken-1.04. See > /home/jkeenan/.cpanm/work/1490881305.4413/build.log for details. > ##### > > However, Task::Weaken succeeded on the second run -- where I was working > with a more recent blead and with App-cpanminus v1.7042 -- and was therefore > already installed in time for the third run. > > So, the only thing I can offer is: Are you working with (a) a recent enough > blead to capture all the changes relevant to no-dot; and (b) the very most > recent version of cpanm?
What you're telling me is you have set your tools up to hide the bug, instead of isolating and reporting the bug. cpanm hides the bug by setting PERL_USE_UNSAFE_INC=1 This is useful for testing for what end users using the cpan toolchains will see, but it is not "The whole problem" This is still a problem outside this subset of ways of doing things. That makes your configuration useful for exposing non-dot-in-inc related issues. But practically useless at exposing dot-in-inc-issues. Subsequently, I don't consider Module::Runtime or Task::Weaken "Fixed" by having tools fake that I have '.' in @INC -- Kent KENTNL - https://metacpan.org/author/KENTNL
