I'm happy to diagnose any false positives. As far as I know everything
is standard, I install using perlbrew, use minicpan to maintain my own
copy and also to see the updates to smoke which are smoked via perl
-MCPAN -e 'CPAN::Shell->test(...'. I don't tweak code or modules.
So with that starting point, let me know what questions you may have and
we can set to diagnosing anything that may be wrong.
On 07/01/2016 05:05 PM, Karen Etheridge wrote:
Hi Nigel,
I want to bring to your attention some serious issues with your smoker
systems that you have been using to submit CPAN Testers reports:
1. there are a very large number of reports (I've received hundreds
from you to my distributions alone, and others have received a large
quantity as well) where prerequisites that are very clearly specified
in Makefile.PL and META.* are not being satisfied. Here is one recent
example:
http://www.cpantesters.org/cpan/report/3cb1d41e-3d45-11e6-9fc5-14bfbd15da07
There was a thread about similar issues on the last year; perhaps you
are experiencing one of the itemized problems as well? (YAML, autodie,
etc)
http://www.nntp.perl.org/group/perl.cpan.testers.discuss/2015/06/msg3650.html
2. the dynamic_config flag in metadata does not appear to be being
respected, e.g. in this report:
http://www.cpantesters.org/cpan/report/9584f7d0-3f0f-11e6-9fc5-14bfbd15da07
The tests are failing because a dynamic prereq has not been
satisfied. The t/00-report-prereqs.t test indicates that MYMETA.json
was never generated, and only static prereqs were ever read. This is
incorrect - dynamic_config => 1 in metadata indicates that Makefile.PL
*must be run* to generate MYMETA.json first, and then that file is
considered authoritative with respect to prereqs.
Hopefully there is enough information in these reports to diagnose the
root issue; if not, no doubt the collective wisdom exists on this
list. thank you!
