Re: [PATCH] ntdll/tests: Skip RtlRandom tests on Windows 10.

[Austin English]

On Mon, Sep 14, 2015 at 12:24 PM, Lauri Kenttä  wrote:
> Windows 10 uses a different algorithm for RtlRandom.
> All the tests currently fail and produce tons of spam.
> (There's already a similar case for Windows Vista.)
>
> dlls/ntdll/tests/rtl.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
>
> diff --git a/dlls/ntdll/tests/rtl.c b/dlls/ntdll/tests/rtl.c
> index 9a4ec48..f42e313 100644
> --- a/dlls/ntdll/tests/rtl.c
> +++ b/dlls/ntdll/tests/rtl.c
> @@ -739,6 +739,16 @@ static void test_RtlRandom(void)
>          return;
>      }
>
> +/*
> -
> - Windows 10 uses different algorithms, so skip the rest of the tests
> -
> - until that is figured out. Trace output for the failures is about 23 MB!
> - */
> -
> - if (seed == 0x7fffffc3) {
> -
>    skip("Most likely running on Windows 10 which uses a different algorithm\n");
> -
>    return;
> - }
> - ok(result == result_expected,
>        "pRtlRandom(&seed (seed == 0)) returns %x, expected %x\n",
>        result, result_expected);
>
> --
> 2.5.1

This should probably be win_skip() rather than skip().

[PATCH] ntdll/tests: Skip RtlRandom tests on Windows 10.

[Lauri Kenttä]

On 2015-09-14 20:28, Austin English wrote:
> On Mon, Sep 14, 2015 at 12:24 PM, Lauri Kenttä  
> wrote:
> > Windows 10 uses a different algorithm for RtlRandom.
> > All the tests currently fail and produce tons of spam.
> > (There's already a similar case for Windows Vista.)
> >
> > dlls/ntdll/tests/rtl.c | 10 ++++++++++
> >  1 file changed, 10 insertions(+)
> >
> > diff --git a/dlls/ntdll/tests/rtl.c b/dlls/ntdll/tests/rtl.c
> > index 9a4ec48..f42e313 100644
> > --- a/dlls/ntdll/tests/rtl.c
> > +++ b/dlls/ntdll/tests/rtl.c
> > @@ -739,6 +739,16 @@ static void test_RtlRandom(void)
> >          return;
> >      }
> >
> > +/*
> > -
> > - Windows 10 uses different algorithms, so skip the rest of the
> >
> > tests
> > -
> > - until that is figured out. Trace output for the failures is about
> >
> > 23 MB!
> > - */
> > -
> > - if (seed == 0x7fffffc3) {
> > -
> >    skip("Most likely running on Windows 10 which uses a 
> >
> > different algorithm\n");
> > -
> >    return;
> > - }
> > - ok(result == result_expected,
> >        "pRtlRandom(&seed (seed == 0)) returns %x, expected %x\n",
> >        result, result_expected);
> >
> > --
> > 2.5.1
>
> This should probably be win_skip() rather than skip().

On the other hand, if Windows is allowed to do that, why should it be an 
error for Wine?

I'm not sure if there's any point in testing the exact output of a PRNG 
at all, especially now that even Windows has multiple versions of it.

[PATCH] ntdll/tests: Skip RtlRandom tests on Windows 10.

[Alexandre Julliard]

Lauri Kenttä  writes:
> On 2015-09-14 20:28, Austin English wrote:
> > This should probably be win_skip() rather than skip().
>
> On the other hand, if Windows is allowed to do that, why should it be
> an error for Wine?
>
> I'm not sure if there's any point in testing the exact output of a
> PRNG at all, especially now that even Windows has multiple versions of
> it.

Yes, the whole thing should probably be replaced by a simple test that
only verifies that we get a different value on every call or something
like that.