Re: d3drm/tests: Release two retrieved interfaces to allow proper cleanup.

[Stefan Dösinger]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Am 2015-10-15 um 21:54 schrieb Bernhard Übelacker:
> - hr = IDirect3DRMFrame_DeleteChild(pFrameP2, pFrameC);
> - ok(hr == D3DRM_OK, "Cannot remove child frame (hr = %x)\n", hr);
> - CHECK_REFCOUNT(pFrameC, 1);
> -
> - CHECK_REFCOUNT(pFrameP1, 2);
> - hr = IDirect3DRMFrame_DeleteChild(pFrameP2, pFrameP1);
> - ok(hr == D3DRM_OK, "Cannot remove child frame (hr = %x)\n", hr);
> - CHECK_REFCOUNT(pFrameP1, 1);

Are these necessary? I'd expect the frame children to be detached when
their parent is destroyed, similarly to how it works in ddraw.

I'm not opposed to explicitly removing them, so I'll send my
signed-off-by to wine-patches. But it's something someone (Jamm? ;-) )
could write an explicit test for.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=R0Nv
-----END PGP SIGNATURE-----

d3drm/tests: Release two retrieved interfaces to allow proper cleanup.

[Bernhard Übelacker]

Am 16.10.2015 um 13:00 schrieb Stefan Dösinger:
> Are these necessary? I'd expect the frame children to be detached when
> their parent is destroyed, similarly to how it works in ddraw.

Thanks for your time.
You are right. To get these released the explicit DeleteChild is not 
neccessary. (I first added these but later failed to realize that I do 
not need them anymore.)

Should I send a changed patch for it?

Kind regards,
Bernhard

d3drm/tests: Release two retrieved interfaces to allow proper cleanup.

[Aaryaman Vasishta]

On Fri, Oct 16, 2015 at 4:30 PM, Stefan Dösinger 
wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
>
> Am 2015-10-15 um 21:54 schrieb Bernhard Übelacker:
> > - hr = IDirect3DRMFrame_DeleteChild(pFrameP2, pFrameC);
> > - ok(hr == D3DRM_OK, "Cannot remove child frame (hr = %x)\n", hr);
> > - CHECK_REFCOUNT(pFrameC, 1);
> > -
> > - CHECK_REFCOUNT(pFrameP1, 2);
> > - hr = IDirect3DRMFrame_DeleteChild(pFrameP2, pFrameP1);
> > - ok(hr == D3DRM_OK, "Cannot remove child frame (hr = %x)\n", hr);
> > - CHECK_REFCOUNT(pFrameP1, 1);
>
> Are these necessary? I'd expect the frame children to be detached when
> their parent is destroyed, similarly to how it works in ddraw.
>
> I'm not opposed to explicitly removing them, so I'll send my
> signed-off-by to wine-patches. But it's something someone (Jamm? ;-) )
> could write an explicit test for.

Well I agree that it's not really needed since the Release method already
deletes any existing children that exist on the frame that it's called on.

d3drm/tests: Release two retrieved interfaces to allow proper cleanup.

[Aaryaman Vasishta]

Then again, I didn't check the refcounting behavior of frames with
children. Maybe these tests could probably be merged with the refcount
tests in test_frame_qi?

Cheers,
Aaryaman

On Sat, Oct 17, 2015 at 12:03 AM, Aaryaman Vasishta <
[e-mail]> wrote:
> On Fri, Oct 16, 2015 at 4:30 PM, Stefan Dösinger <
> [e-mail]> wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA256
> >
> > Am 2015-10-15 um 21:54 schrieb Bernhard Übelacker:
> > > - hr = IDirect3DRMFrame_DeleteChild(pFrameP2, pFrameC);
> > > - ok(hr == D3DRM_OK, "Cannot remove child frame (hr = %x)\n", hr);
> > > - CHECK_REFCOUNT(pFrameC, 1);
> > > -
> > > - CHECK_REFCOUNT(pFrameP1, 2);
> > > - hr = IDirect3DRMFrame_DeleteChild(pFrameP2, pFrameP1);
> > > - ok(hr == D3DRM_OK, "Cannot remove child frame (hr = %x)\n", hr);
> > > - CHECK_REFCOUNT(pFrameP1, 1);
> >
> > Are these necessary? I'd expect the frame children to be detached when
> > their parent is destroyed, similarly to how it works in ddraw.
> >
> > I'm not opposed to explicitly removing them, so I'll send my
> > signed-off-by to wine-patches. But it's something someone (Jamm? ;-) )
> > could write an explicit test for.
>
> Well I agree that it's not really needed since the Release method already
> deletes any existing children that exist on the frame that it's called on.