Re: kernel32: WideCharToMultiByte: return error on negative dest len

[Alexandre Julliard]

"Dan Kegel" <dank(a)kegel.com> writes:
> +static void test_negative_dest_length(void)
> +{
> +    int len;
> +    char buf[10];
> +
> +    /* Test return on negative dest length */
> +    SetLastError( 0xdeadbeef );
> +    memset(buf,'x',sizeof(buf));
> +    len = WideCharToMultiByte(CP_ACP, 0, foobarW, -1, buf, -1, NULL, NULL);
> +    ok(len == 0 && GetLastError() == ERROR_INVALID_PARAMETER,
> +       "WideCharToMultiByte(destlen -1): len=%d error=%x\n", len, GetLastError());

If you want to determine the behavior for negative values you have to
test more than just -1, it may well be special.

-- 
Alexandre Julliard
julliard(a)winehq.org

kernel32: WideCharToMultiByte: return error on negative dest len

[Maarten Lankhorst]

Hello Dan and AJ,

2008/3/24, Alexandre Julliard <julliard(a)winehq.org>:
> "Dan Kegel" <dank(a)kegel.com> writes:
> > +static void test_negative_dest_length(void)
> > +{
> > +    int len;
> > +    char buf[10];
> > +
> > +    /* Test return on negative dest length */
> > +    SetLastError( 0xdeadbeef );
> > +    memset(buf,'x',sizeof(buf));
> > +    len = WideCharToMultiByte(CP_ACP, 0, foobarW, -1, buf, -1, NULL, NULL);
> > +    ok(len == 0 && GetLastError() == ERROR_INVALID_PARAMETER,
> > +       "WideCharToMultiByte(destlen -1): len=%d error=%x\n", len, GetLastError());
>
> If you want to determine the behavior for negative values you have to
>  test more than just -1, it may well be special.
> The test is valid at least on xp, however it could be that wine
> internally relies on the broken behavior, so I'm not 100% sure whether
> it can be committed without checking for that first?

Cheers,
Maarten

kernel32: WideCharToMultiByte: return error on negative dest len

[James Hawkins]

On Mon, Mar 24, 2008 at 10:54 AM, Maarten Lankhorst
<m.b.lankhorst(a)gmail.com> wrote:
> Hello Dan and AJ,
>
> 2008/3/24, Alexandre Julliard <julliard(a)winehq.org>:
> > "Dan Kegel" <dank(a)kegel.com> writes:
> > > +static void test_negative_dest_length(void)
> > > +{
> > > +    int len;
> > > +    char buf[10];
> > > +
> > > +    /* Test return on negative dest length */
> > > +    SetLastError( 0xdeadbeef );
> > > +    memset(buf,'x',sizeof(buf));
> > > +    len = WideCharToMultiByte(CP_ACP, 0, foobarW, -1, buf, -1, NULL, NULL);
> > > +    ok(len == 0 && GetLastError() == ERROR_INVALID_PARAMETER,
> > > +       "WideCharToMultiByte(destlen -1): len=%d error=%x\n", len, GetLastError());
> >
> > If you want to determine the behavior for negative values you have to
> >  test more than just -1, it may well be special.
> >  The test is valid at least on xp, however it could be that wine
> >  internally relies on the broken behavior, so I'm not 100% sure whether
> >  it can be committed without checking for that first?

It's invalid because -1 may be a special case, as Alexandre stated.
That is, there may be a difference between length -1 and length -x,
where x is a positive integer > 1.

-- 
James Hawkins