Thanks Kent. I think it is frustrating for everybody: I can't imagine anyone that would run on to all the troubles to configure a smoker only to generate garbage instead of reliable tests.
There were days that this smoker was generating ~500 reports. As I stated before, it would be quite easily to know something was fishy if I got ~90% of them indicating failure, but that's not the reality. I could also check aleatory samples of the reports to check. What would indicate that reports with failures are presenting problems due smoker issues? Let's use this report from Karen as an example: http://www.cpantesters.org/cpan/report/70299f0a-4f8f-11e6-af52-86b6688db7f4. The only thing that comes to my mind is check for core files, since it seems all that are failing on the smoker are generating core dumps. That includes Moose, for example. I removed everything under the non-root user home directory and started over again. I'm using the standard perl from OpenBSD 5.9 and local::lib, which is the only thing I see that is unusual compared to all other smokers I have created in the past. All other things seems to be OpenBSD specifics, like using a MFS partition for the build_dir option on CPAN, which in fact leads me to ask what the report shows that @INC includes paths from there, considering that I configured CPAN to install everything that is required/recommended as passes the tests: @INC: /var/cpan/MooseX-UndefTolerant-0.20-0/blib/arch /var/cpan/MooseX-UndefTolerant-0.20-0/blib/lib /home/arfreitas/perl5/lib/perl5/5.20.2/amd64-openbsd /home/arfreitas/perl5/lib/perl5/5.20.2 /home/arfreitas/perl5/lib/perl5/amd64-openbsd /home/arfreitas/perl5/lib/perl5 /usr/local/libdata/perl5/site_perl/amd64-openbsd /usr/libdata/perl5/site_perl/amd64-openbsd /usr/local/libdata/perl5/site_perl /usr/libdata/perl5/site_perl /usr/libdata/perl5/amd64-openbsd/5.20.2 /usr/local/libdata/perl5/amd64-openbsd/5.20.2 /usr/libdata/perl5 /usr/local/libdata/perl5 . I can share my notes about smoker preparation if you're willing to check. De: Kent Fredric <kentfred...@gmail.com> Para: Alceu R. de Freitas Jr. <glasswal...@yahoo.com.br> Cc: Karen Etheridge <p...@froods.org>; Alceu R. de Freitas Jr. via Cpan-testers-discuss <cpan-testers-discuss@perl.org> Enviadas: Sábado, 3 de Setembro de 2016 22:47 Assunto: Re: Your smoke testing reports On 4 September 2016 at 11:39, Alceu R. de Freitas Jr. via cpan-testers-discuss <cpan-testers-discuss@perl.org> wrote: > > I'm willing to fix the smoker as long as I got a positive feedback (like > yours). On the other hand, "turn off your smoker or block my id" hardly > qualifies as positive. Smokers of all kinds I just want to say I appreciate. But I really do also understand the frustration of having to wade through lots of reports from smoke boxes that turn out to be problems outside a scope I can even diagnose, be they tooling bugs or smoker misconfiguration. I guess there's some kind of (perhaps unwarrented) expectation that people running smoke boxes should keep an eye on what they're sending to eliminate cases that are obviously "their problem". But eh, I don't know how viable that is for everyone. There's just a shortage of tools that help us here that don't turn into pain for somebody. -- Kent KENTNL - https://metacpan.org/author/KENTNL
