Re: POSIX::strftime() and Unicode [was: RE: Smoke [5.11.2] v5.11.2-81-g162177c FAIL(F) MSWin32 WinXP/.Net SP3 (x86/2 cpu)]

[Nicholas Clark]

On Fri, Nov 27, 2009 at 08:22:38AM -0000, Steve Hay wrote:
> Steve Hay wrote on 2009-11-27:
> > Automated smoke report for 5.11.2 patch
> > 162177c1aed1991639f7f0da64e918c034e1148a v5.11.2-81-g162177c
> > maldoror.bath.planit.group: Intel(R) Core(TM)2 CPU 6700 @
> 2.66GHz(~2660
> > MHz) (x86/2 cpu)
> > on MSWin32 - WinXP/.Net SP3
> > using cl version 14.00.50727.762
> > smoketime 7 hours 25 minutes (average 44 minutes 36 seconds)
> >
> > Summary: FAIL(F)
> >
> > ../ext/POSIX/t/time.t.......................................FAILED
> > 6
>
> This seems to have failed in every configuration because my machine is
> currently booted in Japanese. Normally (in English) it works fine.
>
> I tried the test manually: the failure output is attached (in the hope
> that it is less likely to get mangled than if I copied and pasted it
> here).

Content-Description: out.txt
>
> # Failed test 'strftime() can handle unicode chars in the format string'
> # at t/time.t line 50.
> Wide character in print at ../../lib/Test/Builder.pm line 1698.
> # got: '1970???01???%d???'
> # expected: '1970???01???16???'
> # Looks like you failed 1 test of 13.
> ../ext/POSIX/t/time.t ..

I see just above that test:

# Careful! strftime() is locale sensative. Let's take care of that
my $orig_loc = setlocale(LC_TIME, "C") || die "Cannot setlocale() to C: $!";

Does Windows setlocale() return non-failure, but actually fail to do anything?

(Hence strftime() is still running in a Japanese locale)

Nicholas Clark