Re: [PATCH 2/5] d3drm: Add stubs for IDirect3DRMFrame interface. (v5)
On 19 June 2016 at 00:29, Aaryaman Vasishta <jem456.vasishta(a)gmail.com> wrote:
+static HRESULT WINAPI d3drm_frame1_AddChild(IDirect3DRMFrame *iface, IDirect3DRMFrame *child) +{ + struct d3drm_frame *frame = impl_from_IDirect3DRMFrame(iface); + struct d3drm_frame *child_frame = unsafe_impl_from_IDirect3DRMFrame(child); + + TRACE("iface %p, child %p.\n", iface, child); + + if (!child_frame) + return D3DRMERR_BADOBJECT; That's redundant, d3drm_frame3_AddChild() already takes care of that.
On Sunday, June 19, 2016, Henri Verbeet <hverbeet(a)gmail.com> wrote:
On 19 June 2016 at 00:29, Aaryaman Vasishta <jem456.vasishta(a)gmail.com <javascript:;>> wrote:
+static HRESULT WINAPI d3drm_frame1_AddChild(IDirect3DRMFrame *iface, IDirect3DRMFrame *child) +{ + struct d3drm_frame *frame = impl_from_IDirect3DRMFrame(iface); + struct d3drm_frame *child_frame = unsafe_impl_from_IDirect3DRMFrame(child); + + TRACE("iface %p, child %p.\n", iface, child); + + if (!child_frame) + return D3DRMERR_BADOBJECT; That's redundant, d3drm_frame3_AddChild() already takes care of that.
But child_frame is dereferenced while calling d3drm_frame3_AddChild, so it crashes the test which passes child as NULL. The NULL check for child is made in unsafe_impl_from_... Do correct me if I may have missed something though.
Cheers, Aaryaman
There's a few test failures happening here because of QI and GetParent not being updated for version 1. I am resending this patch along with other changes. Sorry for not catching it earlier :/ On Sun, Jun 19, 2016 at 10:23 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote:
On 19 June 2016 at 18:15, Aaryaman Vasishta <jem456.vasishta(a)gmail.com> wrote:
But child_frame is dereferenced while calling d3drm_frame3_AddChild, Right.
On 19 June 2016 at 21:21, Aaryaman Vasishta <jem456.vasishta(a)gmail.com> wrote:
There's a few test failures happening here because of QI and GetParent not being updated for version 1. I am resending this patch along with other changes. Sorry for not catching it earlier :/
Which tests are that? The changes make sense, but the previous version passed the tests for me.
On Mon, Jun 20, 2016 at 2:07 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote:
On 19 June 2016 at 21:21, Aaryaman Vasishta <jem456.vasishta(a)gmail.com> wrote:
There's a few test failures happening here because of QI and GetParent not being updated for version 1. I am resending this patch along with other changes. Sorry for not catching it earlier :/
Which tests are that? The changes make sense, but the previous version passed the tests for me.
The GetParent tests for version 1 frames, and the QI tests afterwards for the same. The previous version included the updated GetParent and QI, but it slipped by after I added the stubs on top of the rearranged v2 functions. I have sent a v6 of this patch on wine-patches, it will show the changes that I meant.
On Mon, Jun 20, 2016 at 3:11 PM, Aaryaman Vasishta < jem456.vasishta(a)gmail.com> wrote:
On Mon, Jun 20, 2016 at 2:07 PM, Henri Verbeet <hverbeet(a)gmail.com> wrote:
On 19 June 2016 at 21:21, Aaryaman Vasishta <jem456.vasishta(a)gmail.com> wrote:
There's a few test failures happening here because of QI and GetParent not being updated for version 1. I am resending this patch along with other changes. Sorry for not catching it earlier :/
Which tests are that? The changes make sense, but the previous version passed the tests for me.
The GetParent tests for version 1 frames, and the QI tests afterwards for the same. The previous version included the updated GetParent and QI, but it slipped by after I added the stubs on top of the rearranged v2 functions. I have sent a v6 of this patch on wine-patches, it will show the changes that I meant.
Specifically. pFrameTmp == pFrameC. pFrameTmp == pFrameP1, pFrameTmp ==
pFrameP2 tests etc are failing. Also, some vtable tests in test_frame_qi, after they're applied, fail as well. V6 of this patch fixes them.
participants (2)
-
Aaryaman Vasishta -
Henri Verbeet