Re: [ntdll][3/3] Fix returncode for NtDeleteKey (with test)

[Alexandre Julliard]

Paul Vriens  writes:
> Some original tests still fail for NT4, haven't looked into that (yet).
>
> Changelog
>   Fix returncode for NtDeleteKey (with test)

This should be fixed in the server.

[ntdll][3/3] Fix returncode for NtDeleteKey (with test)

[Paul Vriens]

On Wed, 2006-07-05 at 12:57 +0200, Alexandre Julliard wrote:
> Paul Vriens  writes:
> > Some original tests still fail for NT4, haven't looked into that (yet).
> >
> > Changelog
> >   Fix returncode for NtDeleteKey (with test)
>
> This should be fixed in the server.

Will fix it in the server, but can you explain why?

Cheers,

Paul.

[ntdll][3/3] Fix returncode for NtDeleteKey (with test)

[Alexandre Julliard]

Paul Vriens  writes:
> On Wed, 2006-07-05 at 12:57 +0200, Alexandre Julliard wrote:
> > Paul Vriens  writes:
> > > Some original tests still fail for NT4, haven't looked into that (yet).
> > >
> > > Changelog
> > >   Fix returncode for NtDeleteKey (with test)
> >
> > This should be fixed in the server.
>
> Will fix it in the server, but can you explain why?

The server needs to check the handle anyway, there's no reason to do
the check twice.

[ntdll][3/3] Fix returncode for NtDeleteKey (with test)

[Paul Vriens]

On Thu, 2006-07-06 at 10:35 +0200, Alexandre Julliard wrote:
> Paul Vriens  writes:
> > On Wed, 2006-07-05 at 12:57 +0200, Alexandre Julliard wrote:
> > > Paul Vriens  writes:
> > > > Some original tests still fail for NT4, haven't looked into that (yet).
> > > >
> > > > Changelog
> > > >   Fix returncode for NtDeleteKey (with test)
> > >
> > > This should be fixed in the server.
> >
> > Will fix it in the server, but can you explain why?
>
> The server needs to check the handle anyway, there's no reason to do
> the check twice.

Why twice? I return right after the NULL check.

Paul.

[ntdll][3/3] Fix returncode for NtDeleteKey (with test)

[Alexandre Julliard]

Paul Vriens  writes:
> On Thu, 2006-07-06 at 10:35 +0200, Alexandre Julliard wrote:
> > The server needs to check the handle anyway, there's no reason to do
> > the check twice.
>
> Why twice? I return right after the NULL check.

For a NULL handle yes, but that's the should-never-happen case. The
normal case is a valid handle, and then you are checking it twice.

[ntdll][3/3] Fix returncode for NtDeleteKey (with test)

[Joris Huizer]

Paul Vriens wrote:
> > The server needs to check the handle anyway, there's no reason to do
> > the check twice.
>
> Why twice? I return right after the NULL check.
>
> Paul.

I think he meant, if there is a NULL check in the server code, the NULL 
check gets executed twice if it isn't NULL; the first one could be 
avoided then

HTH,

Joris