parallel testing on Win32 (was Re: Profiling parallel tests)
Nicholas Clark
nick at ccl4.org
Wed Sep 2 16:14:14 BST 2009
On Thu, Oct 02, 2008 at 08:31:18PM +0100, Andy Armstrong wrote:
> On 2 Oct 2008, at 14:33, Nicholas Clark wrote:
> >The appended is rather sick, but I'm wondering if it works as a
> >proof of
> >concept. With it, do tests actually run in parallel on Win32?
> >If so, we can work on making it clean.
>
> Hmm. It doesn't seem to work - looks like it's causing results to be
> dropped somewhere. Will investigate further.
Nothing happened on this:
> >Nicholas Clark
> >
> >==== //depot/perl/ext/Test/Harness/lib/TAP/Parser/Iterator/
> >Process.pm#2 - /home/nclark/p4perl/perl/ext/Test/Harness/lib/TAP/
> >Parser/Iterator/Process.pm ====
> >--- /tmp/tmp.11496.95 2008-10-02 14:28:13.000000000 +0100
> >+++ /home/nclark/p4perl/perl/ext/Test/Harness/lib/TAP/Parser/
> >Iterator/Process.pm 2008-10-02 14:06:45.000000000 +0100
> >@@ -95,6 +95,19 @@ sub _use_open3 {
> > eval "use $module";
> > return if $@;
> > }
> >+
> >+ *IPC::Open3::xpipe = sub {
> >+ use Socket qw(AF_UNIX SOCK_STREAM PF_UNSPEC);
> >+
> >+ my $out_w = IO::Handle->new;
> >+ socketpair $_[0], $_[1], AF_UNIX, SOCK_STREAM, PF_UNSPEC
> >+ or die "socketpair: $!";
> >+ # stop reading on the write handle:
> >+ shutdown $_[1], 0 or die "shutdown: $!";
> >+ # stop writing on the read handle:
> >+ shutdown $_[0], 1 or die "shutdown: $!";
> >+ };
> >+
> > return 1;
> >}
> >
>
> --
> Andy Armstrong, Hexten
did it?
Nicholas Clark
More information about the tapx-dev
mailing list