I believe there are two things happening: 1. Nigel's smoker is configured with halt_on_failure => 0 (https://metacpan.org/pod/distribution/CPAN/lib/CPAN/FirstTime.pm#halt_on_failure <https://metacpan.org/pod/distribution/CPAN/lib/CPAN/FirstTime.pm#halt_on_failure>) 2. CPAN::Reporter isn't picking up the failed dependency message and refusing to send in the report. (https://github.com/cpan-testers/CPAN-Reporter/blob/master/lib/CPAN/Reporter.pm#L642-L650 <https://github.com/cpan-testers/CPAN-Reporter/blob/master/lib/CPAN/Reporter.pm#L642-L650>)
(2) assumes that the smoker is using CPAN::Reporter. I don't know if it is or not. Regarding whether there are any 32-bit testers (i686-linux*, i386-freebsd), I suspect there are still, and I could look it up in the database if it's important to know. It looks like stats.cpantesters.org <http://stats.cpantesters.org/> isn't showing more recent versions of Perl since 5.12-ish (which is a bug that should be fixed). Doug Bell d...@preaction.me > On Jul 2, 2018, at 3:48 AM, bulk 88 <bul...@hotmail.com> wrote: > > Here is an interesting report, IO::Compress 2.074 (shipping with core in > 5.25.11 Mar 2017, clearly upgraded for a 5.10) module was installed, > without its prereqs (Compress::Raw::Bzip2 is version 2.020 which is the > version bundled with 5.10.1 Aug 2009) being met, and can't pass a "use > IO::Compress;". > > ----------------------------------------- > Compress::Raw::Bzip2 version 2.074 required--this is only version 2.02 > at > /home/njh/perl5/perlbrew/perls/perl-5.10.1/lib/5.10.1/i686-linux-thread-multi/IO/Compress/Adapter/Bzip2.pm > line 9. > BEGIN failed--compilation aborted at > /home/njh/perl5/perlbrew/perls/perl-5.10.1/lib/5.10.1/i686-linux-thread-multi/IO/Compress/Adapter/Bzip2.pm > line 9. > Compilation failed in require at > /home/njh/perl5/perlbrew/perls/perl-5.10.1/lib/5.10.1/i686-linux-thread-multi/IO/Compress/Bzip2.pm > line 11. > BEGIN failed--compilation aborted at > /home/njh/perl5/perlbrew/perls/perl-5.10.1/lib/5.10.1/i686-linux-thread-multi/IO/Compress/Bzip2.pm > line 11. > Compilation failed in require at DumpDotC.PL line 7. > ----------------------------------------- > > > http://www.cpantesters.org/cpan/report/819555ee-7d79-11e8-8fb1-ef5133556b3f > > IO::Compress, atleast 2.081, does correctly list its dependences for me > ----------------------------------------- > "prereqs" : { > "build" : { > "requires" : { > "ExtUtils::MakeMaker" : "0" > } > }, > "configure" : { > "requires" : { > "ExtUtils::MakeMaker" : "0" > } > }, > "runtime" : { > "requires" : { > "Compress::Raw::Bzip2" : "2.081", > "Compress::Raw::Zlib" : "2.081", > "Scalar::Util" : "0" > } > } > }, > ----------------------------------------- > > So Im wondering how that smoker got IO::Compress tarball installed > without triggering an upgrade of Compress::Raw::Bzip2 tarball. And those > two are always released in lockstep with matching version numbers. > > IDK enough to say, but I dont think the Nigel 5.10.1 smoker ever tested > new IO-Compresses, there is no "4.2.0-1-686-pae" smoker in > http://matrix.cpantesters.org/?dist=IO-Compress%202.081;perl=5.10.1;reports=1 > but there is for my module at > http://matrix.cpantesters.org/?dist=XSConfig%206.26;perl=5.10.1;reports=1 > Its also a coincidence, or is there some bug that IO-Compress has NO > 32 bit test builds for 5.10.1 or is it just the lack of 32 bit testers > on CT anymore? I wonder how new in weeks old the 5.10.1 Nigel smoker is. > I put this on the list incase anyone has ideas how this automatically > happened, unless it was a mistake and manually installed ignoring the > prereqs listed by CPAN shell/Makefile.PL.
signature.asc
Description: Message signed with OpenPGP
