[PATCH v3 4/8] dllhost: Implement IClassFactory::CreateInstance().

[Dmitry Timoshkov]

Signed-off-by: Dmitry Timoshkov <dmitry(a)baikal.ru>
---
 programs/dllhost/dllhost.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/programs/dllhost/dllhost.c b/programs/dllhost/dllhost.c
index 691f233906a..681a20a6bb5 100644
--- a/programs/dllhost/dllhost.c
+++ b/programs/dllhost/dllhost.c
@@ -98,8 +98,14 @@ static ULONG WINAPI factory_Release(IClassFactory *iface)
 static HRESULT WINAPI factory_CreateInstance(IClassFactory *iface,
     IUnknown *punkouter, REFIID iid, void **ppv)
 {
-    FIXME("(%p,%p,%s,%p): stub\n", iface, punkouter, wine_dbgstr_guid(iid), ppv);
-    return E_NOTIMPL;
+    struct factory *factory = impl_from_IClassFactory(iface);
+
+    TRACE("(%p,%p,%s,%p)\n", iface, punkouter, wine_dbgstr_guid(iid), ppv);
+
+    if (punkouter)
+        return CLASS_E_NOAGGREGATION;
+
+    return CoGetClassObject(&factory->clsid, CLSCTX_INPROC_SERVER, NULL, iid, ppv);
 }
 
 static HRESULT WINAPI factory_LockServer(IClassFactory *iface, BOOL lock)
-- 
2.35.1

[Huw Davies]

On Thu, Feb 24, 2022 at 11:50:43AM +0300, Dmitry Timoshkov wrote:
> Signed-off-by: Dmitry Timoshkov <dmitry(a)baikal.ru>
> ---
>  programs/dllhost/dllhost.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/programs/dllhost/dllhost.c b/programs/dllhost/dllhost.c
> index 691f233906a..681a20a6bb5 100644
> --- a/programs/dllhost/dllhost.c
> +++ b/programs/dllhost/dllhost.c
> @@ -98,8 +98,14 @@ static ULONG WINAPI factory_Release(IClassFactory *iface)
>  static HRESULT WINAPI factory_CreateInstance(IClassFactory *iface,
>      IUnknown *punkouter, REFIID iid, void **ppv)
>  {
> -    FIXME("(%p,%p,%s,%p): stub\n", iface, punkouter, wine_dbgstr_guid(iid), ppv);
> -    return E_NOTIMPL;
> +    struct factory *factory = impl_from_IClassFactory(iface);
> +
> +    TRACE("(%p,%p,%s,%p)\n", iface, punkouter, wine_dbgstr_guid(iid), ppv);
> +
> +    if (punkouter)
> +        return CLASS_E_NOAGGREGATION;
> +
> +    return CoGetClassObject(&factory->clsid, CLSCTX_INPROC_SERVER, NULL, iid, ppv);

Shouldn't this ask for IID_IClassFactory, from which CreateInstance is
called?  You can also pass punkouter to the inproc CreateInstance; see
the link you sent me:

https://docs.microsoft.com/en-us/windows/win32/com/writing-a-custom-surrogat...

Huw.

[Dmitry Timoshkov]

Hi Huw,

thanks for the review!

Huw Davies <huw(a)codeweavers.com> wrote:
> > diff --git a/programs/dllhost/dllhost.c b/programs/dllhost/dllhost.c
> > index 691f233906a..681a20a6bb5 100644
> > --- a/programs/dllhost/dllhost.c
> > +++ b/programs/dllhost/dllhost.c
> > @@ -98,8 +98,14 @@ static ULONG WINAPI factory_Release(IClassFactory *iface)
> >  static HRESULT WINAPI factory_CreateInstance(IClassFactory *iface,
> >      IUnknown *punkouter, REFIID iid, void **ppv)
> >  {
> > -    FIXME("(%p,%p,%s,%p): stub\n", iface, punkouter, wine_dbgstr_guid(iid), ppv);
> > -    return E_NOTIMPL;
> > +    struct factory *factory = impl_from_IClassFactory(iface);
> > +
> > +    TRACE("(%p,%p,%s,%p)\n", iface, punkouter, wine_dbgstr_guid(iid), ppv);
> > +
> > +    if (punkouter)
> > +        return CLASS_E_NOAGGREGATION;
> > +
> > +    return CoGetClassObject(&factory->clsid, CLSCTX_INPROC_SERVER, NULL, iid, ppv);
>
> Shouldn't this ask for IID_IClassFactory, from which CreateInstance is
> called?  You can also pass punkouter to the inproc CreateInstance; see
> the link you sent me:
>
> https://docs.microsoft.com/en-us/windows/win32/com/writing-a-custom-surrogat...

Actually this patch could be dropped because according to the results I get
with this patch set IClassFactory::CreateInstance() is never called under Wine
during the surrogate invocation, and if really necessary could be implemented
the way you (and MSDN) suggest. Other patches in the series should apply cleanly
without this patch.

-- 
Dmitry.

[Huw Davies]

On Tue, Mar 08, 2022 at 01:10:15PM +0300, Dmitry Timoshkov wrote:
> Huw Davies <huw(a)codeweavers.com> wrote:
> > Shouldn't this ask for IID_IClassFactory, from which CreateInstance is
> > called?  You can also pass punkouter to the inproc CreateInstance; see
> > the link you sent me:
> >
> > https://docs.microsoft.com/en-us/windows/win32/com/writing-a-custom-surrogat...
>
> Actually this patch could be dropped because according to the results I get
> with this patch set IClassFactory::CreateInstance() is never called under Wine
> during the surrogate invocation, and if really necessary could be implemented
> the way you (and MSDN) suggest. Other patches in the series should apply cleanly
> without this patch.

Hi Dmitry,

Does it all go via IMarshal then?

I guess we can leave this out in that case.  Since I've signed off
the first three patches, could you wait for those to go in and then
send 5-8?

Huw.

[Dmitry Timoshkov]

Huw Davies <huw(a)codeweavers.com> wrote:
> On Tue, Mar 08, 2022 at 01:10:15PM +0300, Dmitry Timoshkov wrote:
> > Huw Davies <huw(a)codeweavers.com> wrote:
> > > Shouldn't this ask for IID_IClassFactory, from which CreateInstance is
> > > called?  You can also pass punkouter to the inproc CreateInstance; see
> > > the link you sent me:
> > >
> > > https://docs.microsoft.com/en-us/windows/win32/com/writing-a-custom-surrogat...
> >
> > Actually this patch could be dropped because according to the results I get
> > with this patch set IClassFactory::CreateInstance() is never called under Wine
> > during the surrogate invocation, and if really necessary could be implemented
> > the way you (and MSDN) suggest. Other patches in the series should apply cleanly
> > without this patch.
>
> Hi Dmitry,
>
> Does it all go via IMarshal then?

Looks like that, yes. At least that's true for the application I'm working
on, and everything seems to work just fine.
> I guess we can leave this out in that case.  Since I've signed off
> the first three patches, could you wait for those to go in and then
> send 5-8?

Sure, sound good. Thanks.

-- 
Dmitry.