I hadn’t noticed that version 0.12 wasn’t throwing such erroneous reports – but that *does* seem to be the case.
Thank you for pointing that out.
CPAN-Test-Dummy-Perl5-Make-CompilationFails-0.02 is designed specifically to let me know when a tester has a configuration that is broken in this way – but, of course, I’ve no guarantee that all testers are testing it.
About a year ago I used to get heaps of such false FAIL reports, mainly from Bingos, such as:http://www.cpantesters.org/cpan/report/401558d6-79da-11e4-87e4-de3106a2f088
It’s not clear to me just where to lay the blame for those particular reports. In the end the only way I could avoid those FAILs was to detect the absence of quadmath.h during the Makefile.PL processing and exit immediately.
I suspect that Bingos’ smokers would no longer throw those false FAIL reports, but that Makefile.PL processing (which was only ever performed for NetBSD systems) is still in place anyway.
From: Karen Etheridge
Sent: Monday, March 07, 2016 4:42 AM
Cc: CPAN Testers Discuss
Subject: Re: Avoiding App::cpanminus::reporter
I have been running the previous release of App::cpanminus::reporter (0.12) for quite some time because of this issue: https://rt.cpan.org/Ticket/Display.html?id=103866
I have not run into the situation (as far as I know) where UNKNOWN reports are reported as FAIL, either as a tester or an author, but it's quite possible that I've overlooked these occurrences.
On Sun, Mar 6, 2016 at 4:07 AM, wrote:
As many of us are no doubt aware, App::cpanminus::reporter is critically broken - as reported by Slaven Rezic over 15 months ago in this bug report:
Whenever I get a false FAIL report because of this bug, I've started sending emails to the tester from whom the false FAIL emanated.
It occurs to me that, as an additional courtesy, I should also be telling the tester the name of the module with which to replace App::cpanminus::reporter.
Is CPAN::Reporter the recommended replacement ?