On Wed, 27 Aug 2003, Stas Bekman wrote:
> Stas Bekman wrote:
> > Randy Kobes wrote:
> >> On Wed, 27 Aug 2003, Stas Bekman wrote:
> >>> Randy Kobes wrote:
> >>>> On Mon, 25 Aug 2003, Stas Bekman wrote:
> >>>>> It's possible that the problem comes from the new PerlIO
> >>>>> Layer, try testing again after disabling it:
> >>>> [ .. ]
> >>>> You're right - disabling PERLIO_LAYERS allows the
> >>>> modules/cgi.t and modules/cgiupload.t to pass ...
> >>>> I'll try to track down what changes in the past few
> >>>> weeks led to this. Thanks.
>
> Hmm, is this still true? from t/response/TestAPR/perlio.pm:
>
> # 1) APR::PerlIO is not available on all platforms
> # 2) on those platforms where available,
> # PerlIO-enabled perl is needed to run this test
> my $run = eval { require APR::PerlIO } &&
> APR::PerlIO::PERLIO_LAYERS_ARE_ENABLED()
> ? 1
> : have { "This Perl build doesn't support PerlIO layers" => 0 };
>
> And if so, that explains the problem with moving to IO layers as the IO
> "transport" on win32.
>
> Hasn't this been fixed? I remember the problem was with the missing symbol
> which is fixed for sure. can we remove eval {} there?
Hi Stas,
This has been fixed, and the associated perlio.t test
runs on Win32, and passes all but one subtest (I don't have
the details handy, but the failure is due to a difference
of a single leading space between expected and actual
output).
--
best regards,
randy
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
