Bug #127126 for Net-Async-HTTP: t/91rt100066.t fails against IO::Async::Loop::Epoll

Mon Sep 17 09:59:36 2018 leonerd-cpan [...] leonerd.org.uk - Ticket created

Subject:

t/91rt100066.t fails against IO::Async::Loop::Epoll

$ IO_ASYNC_LOOP=Epoll prove t/91rt100066.t t/91rt100066.t .. 1/? Oopsie at t/91rt100066.t line 41. # Failed test 'Oopsie exception caught at loop toplevel' # at t/91rt100066.t line 59. # undef # doesn't match '(?^:^Oopsie at t\/91rt100066\.t line \d+)' # Looks like you failed 1 test of 9. t/91rt100066.t .. Dubious, test returned 1 (wstat 256, 0x100) Failed 1/9 subtests $ IO_ASYNC_LOOP=Ppoll prove t/91rt100066.t t/91rt100066.t .. ok All tests successful. -- Paul Evans

Mon Sep 17 10:04:10 2018 leonerd-cpan [...] leonerd.org.uk - Correspondence added

The plot thickens: $ perlmodversion IO::Async::Loop::Epoll 0.19 $ perlmodversion Linux::Epoll 0.015 $ IO_ASYNC_LOOP=Epoll prove t/91rt100066.t t/91rt100066.t .. 1/? Oopsie at t/91rt100066.t line 41. ... => FAIL (on another machine) $ perlmodversion IO::Async::Loop::Epoll 0.19 $ perlmodversion Linux::Epoll 0.013 $ IO_ASYNC_LOOP=Epoll prove -b t/91rt100066.t t/91rt100066.t .. ok All tests successful. -- Paul Evans

Mon Sep 17 10:08:57 2018 leonerd-cpan [...] leonerd.org.uk - Correspondence added

On Mon Sep 17 10:04:10 2018, PEVANS wrote: Show quoted text

> (on another machine) > $ perlmodversion IO::Async::Loop::Epoll > 0.19 > > $ perlmodversion Linux::Epoll > 0.013

Updating this machine to Linux::Epoll 0.015 now makes it fail the same way. I suspect something in Linux::Epoll having changed the way that 'die' is handled. -- Paul Evans

Mon Sep 17 10:11:24 2018 leonerd-cpan [...] leonerd.org.uk - Correspondence added

Reported upstream https://rt.cpan.org/Ticket/Display.html?id=127127 -- Paul Evans