Re: [1/2] dimm.idl: Add IActiveIMMMessagePumpOwner.

[Nikolay Sivov]

What's a reason for having it declared like that:

--
+    HRESULT OnTranslateMessage([in] const struct tagMSG *pMsg);
---

Why not use MSG*?

[1/2] dimm.idl: Add IActiveIMMMessagePumpOwner.

[Thomas Faber]

On 2015-06-23 23:51, Nikolay Sivov wrote:
> What's a reason for having it declared like that:
>
> --
> +    HRESULT OnTranslateMessage([in] const struct tagMSG *pMsg);
> ---
>
> Why not use MSG*?

Hmm I thought the compiler was complaining about not knowing MSG, but I
must have imagined that, it's building fine now.
I don't think it really makes a difference, but I can resend with that
fixed in both files if you think it's cleaner.

Thanks,
Thomas

[1/2] dimm.idl: Add IActiveIMMMessagePumpOwner.

[Nikolay Sivov]

On 24.06.2015 0:58, Thomas Faber wrote:
> On 2015-06-23 23:51, Nikolay Sivov wrote:
> > What's a reason for having it declared like that:
> >
> > --
> > +    HRESULT OnTranslateMessage([in] const struct tagMSG *pMsg);
> > ---
> >
> > Why not use MSG*?
>
> Hmm I thought the compiler was complaining about not knowing MSG, but I
> must have imagined that, it's building fine now.
> I don't think it really makes a difference, but I can resend with that
> fixed in both files if you think it's cleaner.

Yes, I think there's no reason not to match SDK in that regard.
> [
> +    object,
> +    uuid(b5cf2cfa-8aeb-11d1-9364-0060b067b86e),
> +    pointer_default(unique),
> +    local
> +]

Why local? SDK idl doesn't have it.
> +    HRESULT OnTranslateMessage([in] const struct tagMSG *pMsg);
> +    HRESULT Pause([out] DWORD *pwdCookie);
> +    HRESULT Resume([in] DWORD dwCookie);

Also please don't use prefixed notation, 'msg' and 'cookie' (in Pause 
and Resume) would be better.
> Thanks,
> Thomas

[1/2] dimm.idl: Add IActiveIMMMessagePumpOwner.

[Thomas Faber]

On 2015-06-24 09:15, Nikolay Sivov wrote:
> On 24.06.2015 0:58, Thomas Faber wrote:
> > On 2015-06-23 23:51, Nikolay Sivov wrote:
> > > What's a reason for having it declared like that:
> > >
> > > --
> > > +    HRESULT OnTranslateMessage([in] const struct tagMSG *pMsg);
> > > ---
> > >
> > > Why not use MSG*?
> >
> > Hmm I thought the compiler was complaining about not knowing MSG, but I
> > must have imagined that, it's building fine now.
> > I don't think it really makes a difference, but I can resend with that
> > fixed in both files if you think it's cleaner.
>
> Yes, I think there's no reason not to match SDK in that regard.

Will do.
> > [
> > +    object,
> > +    uuid(b5cf2cfa-8aeb-11d1-9364-0060b067b86e),
> > +    pointer_default(unique),
> > +    local
> > +]
>
> Why local? SDK idl doesn't have it.
> > +    HRESULT OnTranslateMessage([in] const struct tagMSG *pMsg);
> > +    HRESULT Pause([out] DWORD *pwdCookie);
> > +    HRESULT Resume([in] DWORD dwCookie);
>
> Also please don't use prefixed notation, 'msg' and 'cookie' (in Pause 
> and Resume) would be better.

I was following what the rest of the file does with these two things.
Will change it.


Thanks.
-Thomas